This is the Buildbot documentation for Buildbot version |version|.

If you are evaluating Buildbot and would like to get started quickly, start with the Tutorial. Regular users of Buildbot should consult the Manual, and those wishing to modify Buildbot directly will want to be familiar with the Developer's Documentation.

Table Of Contents

1. Buildbot Tutorial


1.1. First Run

1.1.1. Goal

This tutorial will take you from zero to running your first buildbot master and worker as quickly as possible, without changing the default configuration.

This tutorial is all about instant gratification and the five minute experience: in five minutes we want to convince you that this project Works, and that you should seriously consider spending some more time learning the system. In this tutorial no configuration or code changes are done.

This tutorial assumes that you are running on Unix, but might be adaptable easily to Windows.

Thanks to virtualenv, installing buildbot in a standalone environment is very easy. For those more familiar with Docker, there also exists a docker version of these instructions.

You should be able to cut and paste each shell block from this tutorial directly into a terminal.

1.1.2. Getting ready

There are many ways to get the code on your machine. We will use here the easiest one: via pip in a virtualenv. It has the advantage of not polluting your operating system, as everything will be contained in the virtualenv.

To make this work, you will need the following installed:

Preferably, use your distribution package manager to install these.

You will also need a working Internet connection, as virtualenv and pip will need to download other projects from the Internet.


Buildbot does not require root access. Run the commands in this tutorial as a normal, unprivileged user.

1.1.3. Creating a master

The first necessary step is to create a virtualenv for our master. All our operations will happen in this directory:

mkdir tmp
cd tmp
virtualenv --no-site-packages bb-master
cd bb-master

Now that we are ready, we need to install buildbot:

./bin/pip install buildbot[bundle]

Now that buildbot is installed, it's time to create the master:

./bin/buildbot create-master master

Buildbot's activity is controlled by a configuration file. We will use the sample configuration file unchanged:

mv master/master.cfg.sample master/master.cfg

Finally, start the master:

./bin/buildbot start master

You will now see some log information from the master in this terminal. It should ends with lines like these:

2014-11-01 15:52:55+0100 [-] BuildMaster is running
The buildmaster appears to have (re)started correctly.

From now on, feel free to visit the web status page running on the port 8010: http://localhost:8010/

Our master now needs (at least) a worker to execute its commands. For that, heads on to the next section !

1.1.4. Creating a worker

The worker will be executing the commands sent by the master. In this tutorial, we are using the pyflakes project as an example. As a consequence of this, your worker will need access to the git command in order to checkout some code. Be sure that it is installed, or the builds will fail.

Same as we did for our master, we will create a virtualenv for our worker next to the other one. It would however be completely ok to do this on another computer - as long as the worker computer is able to connect to the master one:

cd tmp
virtualenv --no-site-packages bb-worker
cd bb-worker

Install the buildbot-worker command:

./bin/pip install buildbot-worker

Now, create the worker:

./bin/buildbot-worker create-worker worker localhost example-worker pass


If you decided to create this from another computer, you should replace localhost with the name of the computer where your master is running.

The username (example-worker), and password (pass) should be the same as those in master/master.cfg; verify this is the case by looking at the section for c['workers']:

cat master/master.cfg

And finally, start the worker:

./bin/buildbot-worker start worker

Check the worker's output. It should end with lines like these:

2014-11-01 15:56:51+0100 [-] Connecting to localhost:9989
2014-11-01 15:56:51+0100 [Broker,client] message from master: attached
The worker appears to have (re)started correctly.

Meanwhile, from the other terminal, in the master log (:file:twisted.log in the master directory), you should see lines like these:

2014-11-01 15:56:51+0100 [Broker,1,] worker 'example-worker' attaching from IPv4Address(TCP, '', 54015)
2014-11-01 15:56:51+0100 [Broker,1,] Got workerinfo from 'example-worker'
2014-11-01 15:56:51+0100 [-] bot attached

You should now be able to go to http://localhost:8010, where you will see a web page similar to:

index page

Click on the Waterfall Display link and you get this:

empty waterfall.

Your master is now quietly waiting for new commits to Pyflakes. This doesn't happen very often though. In the next section, we'll see how to manually start a build.

We just wanted to get you to dip your toes in the water. It's easy to take your first steps, but this is about as far as we can go without touching the configuration.

You've got a taste now, but you're probably curious for more. Let's step it up a little in the second tutorial by changing the configuration and doing an actual build. Continue on to A Quick Tour.

1.2. First Buildbot run with Docker


Docker can be tricky to get working correctly if you haven't used it before. If you're having trouble, first determine whether it is a Buildbot issue or a Docker issue by running:

docker run ubuntu:12.04 apt-get update

If that fails, look for help with your Docker install. On the other hand, if that succeeds, then you may have better luck getting help from members of the Buildbot community.

Docker is a tool that makes building and deploying custom environments a breeze. It uses lightweight linux containers (LXC) and performs quickly, making it a great instrument for the testing community. The next section includes a Docker pre-flight check. If it takes more that 3 minutes to get the 'Success' message for you, try the Buildbot pip-based first run instead.

1.2.1. Current Docker dependencies

  • Linux system, with at least kernel 3.8 and AUFS support. For example, Standard Ubuntu, Debian and Arch systems.
  • Packages: lxc, iptables, ca-certificates, and bzip2 packages.
  • Local clock on time or slightly in the future for proper SSL communication.
  • This tutorial uses docker-compose to run a master, a worker, and a postgresql database server

1.2.2. Installation

1.2.3. Building and running Buildbot

# Download Buildbot docker-compose.yml.

# Build the Buildbot container (it will take a few minutes to download packages)
docker-compose up

You should now be able to go to http://localhost:8010 and see a web page similar to:

index page

Click on the Waterfall Display link and you get this:

empty waterfall.

1.2.4. Overview of the docker-compose configuration

This docker-compose configuration is made as a basis for what you would put in production

  • Separated containers for each component
  • A solid database backend with postgresql
  • A buildbot master that exposes its configuration to the docker host
  • A buildbot worker that can be cloned in order to add additional power
  • Containers are linked together so that the only port exposed to external is the web server
  • The default master container is based on Alpine linux for minimal footprint
  • The default worker container is based on more widely known Ubuntu distribution, as this is the container you want to customize.

1.2.5. Playing with your Buildbot containers

If you've come this far, you have a Buildbot environment that you can freely experiment with. The container is storing all its valuable information in /var/lib/buildbot and /var/lib/buildbot_db

You can access your buildbot configuration in /var/lib/buildbot

vi /var/lib/buildbot/master.cfg

1.2.6. Customize your Worker container

It is advised to customize you worker container in order to suit your project's build dependencies and need. An example DockerFile is available in the contrib directory of buildbot:

You've got a taste now, but you're probably curious for more. Let's step it up a little in the second tutorial by changing the configuration and doing an actual build. Continue on to A Quick Tour.

1.3. A Quick Tour

1.3.1. Goal

This tutorial will expand on the First Run tutorial by taking a quick tour around some of the features of buildbot that are hinted at in the comments in the sample configuration. We will simply change parts of the default configuration and explain the activated features.

As a part of this tutorial, we will make buildbot do a few actual builds.

This section will teach you how to:

  • make simple configuration changes and activate them
  • deal with configuration errors
  • force builds
  • enable and control the IRC bot
  • enable ssh debugging
  • add a 'try' scheduler

1.3.2. Setting Project Name and URL

Let's start simple by looking at where you would customize the buildbot's project name and URL.

We continue where we left off in the First Run tutorial.

Open a new terminal, and first enter the same sandbox you created before (where $EDITOR is your editor of choice like vim, gedit, or emacs):

cd tmp/buildbot
source sandbox/bin/activate
$EDITOR master/master.cfg

Now, look for the section marked PROJECT IDENTITY which reads:


# the 'title' string will appear at the top of this buildbot installation's
# home pages (linked to the 'titleURL').

c['title'] = "Pyflakes"
c['titleURL'] = ""

If you want, you can change either of these links to anything you want to see what happens when you change them.

After making a change go into the terminal and type:

buildbot reconfig master

You will see a handful of lines of output from the master log, much like this:

2011-12-04 10:11:09-0600 [-] loading configuration from /home/dustin/tmp/buildbot/master/master.cfg
2011-12-04 10:11:09-0600 [-] configuration update started
2011-12-04 10:11:09-0600 [-] builder runtests is unchanged
2011-12-04 10:11:09-0600 [-] removing IStatusReceiver <WebStatus on port tcp:8010 at 0x2aee368>
2011-12-04 10:11:09-0600 [-] (TCP Port 8010 Closed)
2011-12-04 10:11:09-0600 [-] Stopping factory <buildbot.status.web.baseweb.RotateLogSite instance at 0x2e36638>
2011-12-04 10:11:09-0600 [-] adding IStatusReceiver <WebStatus on port tcp:8010 at 0x2c2d950>
2011-12-04 10:11:09-0600 [-] RotateLogSite starting on 8010
2011-12-04 10:11:09-0600 [-] Starting factory <buildbot.status.web.baseweb.RotateLogSite instance at 0x2e36e18>
2011-12-04 10:11:09-0600 [-] Setting up http.log rotating 10 files of 10000000 bytes each
2011-12-04 10:11:09-0600 [-] WebStatus using (/home/dustin/tmp/buildbot/master/public_html)
2011-12-04 10:11:09-0600 [-] removing 0 old schedulers, updating 0, and adding 0
2011-12-04 10:11:09-0600 [-] adding 1 new changesources, removing 1
2011-12-04 10:11:09-0600 [-] gitpoller: using workdir '/home/dustin/tmp/buildbot/master/gitpoller-workdir'
2011-12-04 10:11:09-0600 [-] GitPoller repository already exists
2011-12-04 10:11:09-0600 [-] configuration update complete

Reconfiguration appears to have completed successfully.

The important lines are the ones telling you that it is loading the new configuration at the top, and the one at the bottom saying that the update is complete.

Now, if you go back to the waterfall page, you will see that the project's name is whatever you may have changed it to and when you click on the URL of the project name at the bottom of the page it should take you to the link you put in the configuration.

1.3.3. Configuration Errors

It is very common to make a mistake when configuring buildbot, so you might as well see now what happens in that case and what you can do to fix the error.

Open up the config again and introduce a syntax error by removing the first single quote in the two lines you changed, so they read:

c[title'] = "Pyflakes
c[titleURL'] = ""

This creates a Python SyntaxError. Now go ahead and reconfig the buildmaster:

buildbot reconfig master

This time, the output looks like:

2015-08-14 18:40:46+0000 [-] beginning configuration update
2015-08-14 18:40:46+0000 [-] Loading configuration from '/data/buildbot/master/master.cfg'
2015-08-14 18:40:46+0000 [-] error while parsing config file:
        Traceback (most recent call last):
          File "/usr/local/lib/python2.7/dist-packages/buildbot/", line 265, in reconfig
            d = self.doReconfig()
          File "/usr/local/lib/python2.7/dist-packages/twisted/internet/", line 1274, in unwindGenerator
            return _inlineCallbacks(None, gen, Deferred())
          File "/usr/local/lib/python2.7/dist-packages/twisted/internet/", line 1128, in _inlineCallbacks
            result = g.send(result)
          File "/usr/local/lib/python2.7/dist-packages/buildbot/", line 289, in doReconfig
        --- <exception caught here> ---
          File "/usr/local/lib/python2.7/dist-packages/buildbot/", line 156, in loadConfig
            exec f in localDict
        exceptions.SyntaxError: EOL while scanning string literal (master.cfg, line 103)

2015-08-14 18:40:46+0000 [-] error while parsing config file: EOL while scanning string literal (master.cfg, line 103) (traceback in logfile)
2015-08-14 18:40:46+0000 [-] reconfig aborted without making any changes

Reconfiguration failed. Please inspect the master.cfg file for errors,
correct them, then try 'buildbot reconfig' again.

This time, it's clear that there was a mistake in the configuration. Luckily, the Buildbot master will ignore the wrong configuration and keep running with the previous configuration.

The message is clear enough, so open the configuration again, fix the error, and reconfig the master.

1.3.4. Your First Build

By now you're probably thinking: "All this time spent and still not done a single build? What was the name of this project again?"

On the waterfall page, click on the runtests link. You'll see a builder page, and in the upper-right corner is a box where you can login. The default username and password are both "pyflakes". Once you've logged in, you will see some new options that allow you to force a build:

force a build.

Click Force Build - there's no need to fill in any of the fields in this case. Next, click on view in waterfall.

You will now see:

an successful test run happened.

1.3.5. Enabling the IRC Bot

Buildbot includes an IRC bot that you can tell to join a channel and control to report on the status of buildbot.

First, start an IRC client of your choice, connect to and join an empty channel. In this example we will use #buildbot-test, so go join that channel. (Note: please do not join the main buildbot channel!)

Edit :file:'master.cfg' and look for the STATUS TARGETS section. At the end of that section add the lines:

from buildbot.status import irc
c['status'].append(irc.IRC(host="", nick="bbtest",

Reconfigure the build master then do:

grep -i irc master/twistd.log

The log output should contain a line like this:

2015-08-14 20:00:33+0000 [-] Starting factory <buildbot.status.words.IrcStatusFactory instance at 0x7fee15c640e0>
2015-08-14 20:00:48+0000 [IrcStatusBot,client] <buildbot.status.words.IrcStatusBot instance at 0x7fee1653f1b8>: I have joined #buildbot-test

You should see the bot now joining in your IRC client. In your IRC channel, type:

bbtest: commands

to get a list of the commands the bot supports.

Let's tell the bot to notify certain events, to learn which EVENTS we can notify on:

bbtest: help notify

Now let's set some event notifications:

bbtest: notify on started
bbtest: notify on finished
bbtest: notify on failure

The bot should have responded to each of the commands:

<@lsblakk> bbtest: notify on started
<bbtest> The following events are being notified: ['started']
<@lsblakk> bbtest: notify on finished
<bbtest> The following events are being notified: ['started', 'finished']
<@lsblakk> bbtest: notify on failure
<bbtest> The following events are being notified: ['started', 'failure', 'finished']

Now, go back to the web interface and force another build.

Notice how the bot tells you about the start and finish of this build:

< bbtest> build #1 of runtests started, including []
< bbtest> build #1 of runtests is complete: Success [build successful]  Build details are at http://localhost:8010/builders/runtests/builds/1

You can also use the bot to force a build:

bbtest: force build runtests test build

But to allow this, you'll need to have allowForce in the IRC configuration:

c['status'].append(irc.IRC(host="", nick="bbtest",

This time, the bot is giving you more output, as it's specifically responding to your direct request to force a build, and explicitly tells you when the build finishes:

<@lsblakk> bbtest: force build runtests test build
< bbtest> build #2 of runtests started, including []
< bbtest> build forced [ETA 0 seconds]
< bbtest> I'll give a shout when the build finishes
< bbtest> build #2 of runtests is complete: Success [build successful]  Build details are at http://localhost:8010/builders/runtests/builds/2

You can also see the new builds in the web interface.

a successful test run from IRC happened.

1.3.6. Setting Authorized Web Users

Further down, look for the WebStatus configuration:

c['status'] = []

from buildbot.status import html
from buildbot.status.web import authz, auth

    # change any of these to True to enable; see the manual for more
    # options
    gracefulShutdown = False,
    forceBuild = 'auth',  # use this to test your worker once it is set up
    forceAllBuilds = False,
    pingBuilder = False,
    stopBuild = False,
    stopAllBuilds = False,
    cancelPendingBuild = False,
c['status'].append(html.WebStatus(http_port=8010, authz=authz_cfg))

The auth.BasicAuth() define authorized users and their passwords. You can change these or add new ones.

1.3.7. Debugging with Manhole

You can do some debugging by using manhole, an interactive Python shell. It exposes full access to the buildmaster's account (including the ability to modify and delete files), so it should not be enabled with a weak or easily guessable password.

To use this you will need to install an additional package or two to your virtualenv:

cd tmp/buildbot
source sandbox/bin/activate
easy_install pycrypto
easy_install pyasn1

In your master.cfg find:

c = BuildmasterConfig = {}

Insert the following to enable debugging mode with manhole:

from buildbot import manhole
c['manhole'] = manhole.PasswordManhole("tcp:1234:interface=","admin","passwd")

After restarting the master, you can ssh into the master and get an interactive Python shell:

ssh -p1234 admin@
# enter passwd at prompt


The pyasn1-0.1.1 release has a bug which results in an exception similar to this on startup:

exceptions.TypeError: argument 2 must be long, not int

If you see this, the temporary solution is to install the previous version of pyasn1:

pip install pyasn1-0.0.13b

If you wanted to check which workers are connected and what builders those workers are assigned to you could do:

>>> master.workers.workers
{'example-worker': <Worker 'example-worker', current builders: runtests>}

Objects can be explored in more depth using dir(x) or the helper function show(x).

1.3.8. Adding a 'try' scheduler

Buildbot includes a way for developers to submit patches for testing without committing them to the source code control system. (This is really handy for projects that support several operating systems or architectures.)

To set this up, add the following lines to master.cfg:

from buildbot.scheduler import Try_Userpass
c['schedulers'] = []

Then you can submit changes using the try command.

Let's try this out by making a one-line change to pyflakes, say, to make it trace the tree by default:

git clone git:// pyflakes-git
cd pyflakes-git/pyflakes
# change "traceTree = False" on line 185 to "traceTree = True"

Then run buildbot's try command as follows:

source ~/tmp/buildbot/sandbox/bin/activate
buildbot try --connect=pb --master= --username=sampleuser --passwd=samplepass --vc=git

This will do git diff for you and send the resulting patch to the server for build and test against the latest sources from Git.

Now go back to the waterfall page, click on the runtests link, and scroll down. You should see that another build has been started with your change (and stdout for the tests should be chock-full of parse trees as a result). The "Reason" for the job will be listed as "'try' job", and the blamelist will be empty.

To make yourself show up as the author of the change, use the --who=emailaddr option on buildbot try to pass your email address.

To make a description of the change show up, use the --properties=comment="this is a comment" option on buildbot try.

To use ssh instead of a private username/password database, see Try_Jobdir.

1.4. Further Reading

See the following user-contributed tutorials for other highlights and ideas:

1.4.1. Buildbot in 5 minutes - a user-contributed tutorial

(Ok, maybe 10.)

Buildbot is really an excellent piece of software, however it can be a bit confusing for a newcomer (like me when I first started looking at it). Typically, at first sight it looks like a bunch of complicated concepts that make no sense and whose relationships with each other are unclear. After some time and some reread, it all slowly starts to be more and more meaningful, until you finally say "oh!" and things start to make sense. Once you get there, you realize that the documentation is great, but only if you already know what it's about.

This is what happened to me, at least. Here I'm going to (try to) explain things in a way that would have helped me more as a newcomer. The approach I'm taking is more or less the reverse of that used by the documentation, that is, I'm going to start from the components that do the actual work (the builders) and go up the chain from there up to change sources. I hope purists will forgive this unorthodoxy. Here I'm trying to clarify the concepts only, and will not go into the details of each object or property; the documentation explains those quite well. Installation

I won't cover the installation; both Buildbot master and worker are available as packages for the major distributions, and in any case the instructions in the official documentation are fine. This document will refer to Buildbot 0.8.5 which was current at the time of writing, but hopefully the concepts are not too different in other versions. All the code shown is of course python code, and has to be included in the master.cfg master configuration file.

We won't cover the basic things such as how to define the workers, project names, or other administrative information that is contained in that file; for that, again the official documentation is fine. Builders: the workhorses

Since Buildbot is a tool whose goal is the automation of software builds, it makes sense to me to start from where we tell Buildbot how to build our software: the builder (or builders, since there can be more than one).

Simply put, a builder is an element that is in charge of performing some action or sequence of actions, normally something related to building software (for example, checking out the source, or make all), but it can also run arbitrary commands.

A builder is configured with a list of workers that it can use to carry out its task. The other fundamental piece of information that a builder needs is, of course, the list of things it has to do (which will normally run on the chosen worker). In Buildbot, this list of things is represented as a BuildFactory object, which is essentially a sequence of steps, each one defining a certain operation or command.

Enough talk, let's see an example. For this example, we are going to assume that our super software project can be built using a simple make all, and there is another target make packages that creates rpm, deb and tgz packages of the binaries. In the real world things are usually more complex (for example there may be a configure step, or multiple targets), but the concepts are the same; it will just be a matter of adding more steps to a builder, or creating multiple builders, although sometimes the resulting builders can be quite complex.

So to perform a manual build of our project we would type this from the command line (assuming we are at the root of the local copy of the repository):

$ make clean    # clean remnants of previous builds
$ svn update
$ make all
$ make packages
# optional but included in the example: copy packages to some central machine
$ scp packages/*.rpm packages/*.deb packages/*.tgz someuser@somehost:/repository

Here we're assuming the repository is SVN, but again the concepts are the same with git, mercurial or any other VCS.

Now, to automate this, we create a builder where each step is one of the commands we typed above. A step can be a shell command object, or a dedicated object that checks out the source code (there are various types for different repositories, see the docs for more info), or yet something else:

from buildbot.plugins import steps, util

# first, let's create the individual step objects

# step 1: make clean; this fails if the worker has no local copy, but
# is harmless and will only happen the first time
makeclean = steps.ShellCommand(name="make clean",
                               command=["make", "clean"],
                               description="make clean")

# step 2: svn update (here updates trunk, see the docs for more
# on how to update a branch, or make it more generic).
checkout = steps.SVN(baseURL='svn://myrepo/projects/coolproject/trunk',

# step 3: make all
makeall = steps.ShellCommand(name="make all",
                             command=["make", "all"],
                             description="make all")

# step 4: make packages
makepackages = steps.ShellCommand(name="make packages",
                                  command=["make", "packages"],
                                  description="make packages")

# step 5: upload packages to central server. This needs passwordless ssh
# from the worker to the server (set it up in advance as part of worker setup)
uploadpackages = steps.ShellCommand(name="upload packages",
                                    description="upload packages",
                                    command="scp packages/*.rpm packages/*.deb packages/*.tgz someuser@somehost:/repository",

# create the build factory and add the steps to it
f_simplebuild = util.BuildFactory()

# finally, declare the list of builders. In this case, we only have one builder
c['builders'] = [
    util.BuilderConfig(name="simplebuild", workernames=['worker1', 'worker2', 'worker3'], factory=f_simplebuild)

So our builder is called simplebuild and can run on either of worker1, worker2 and worker3. If our repository has other branches besides trunk, we could create another one or more builders to build them; in the example, only the checkout step would be different, in that it would need to check out the specific branch. Depending on how exactly those branches have to be built, the shell commands may be recycled, or new ones would have to be created if they are different in the branch. You get the idea. The important thing is that all the builders be named differently and all be added to the c['builders'] value (as can be seen above, it is a list of BuilderConfig objects).

Of course the type and number of steps will vary depending on the goal; for example, to just check that a commit doesn't break the build, we could include just up to the make all step. Or we could have a builder that performs a more thorough test by also doing make test or other targets. You get the idea. Note that at each step except the very first we use haltOnFailure=True because it would not make sense to execute a step if the previous one failed (ok, it wouldn't be needed for the last step, but it's harmless and protects us if one day we add another step after it). Schedulers

Now this is all nice and dandy, but who tells the builder (or builders) to run, and when? This is the job of the scheduler, which is a fancy name for an element that waits for some event to happen, and when it does, based on that information decides whether and when to run a builder (and which one or ones). There can be more than one scheduler. I'm being purposely vague here because the possibilities are almost endless and highly dependent on the actual setup, build purposes, source repository layout and other elements.

So a scheduler needs to be configured with two main pieces of information: on one hand, which events to react to, and on the other hand, which builder or builders to trigger when those events are detected. (It's more complex than that, but if you understand this, you can get the rest of the details from the docs).

A simple type of scheduler may be a periodic scheduler: when a configurable amount of time has passed, run a certain builder (or builders). In our example, that's how we would trigger a build every hour:

from buildbot.plugins import schedulers

# define the periodic scheduler
hourlyscheduler = schedulers.Periodic(name="hourly",

# define the available schedulers
c['schedulers'] = [hourlyscheduler]

That's it. Every hour this hourly scheduler will run the simplebuild builder. If we have more than one builder that we want to run every hour, we can just add them to the builderNames list when defining the scheduler and they will all be run. Or since multiple scheduler are allowed, other schedulers can be defined and added to c['schedulers'] in the same way.

Other types of schedulers exist; in particular, there are schedulers that can be more dynamic than the periodic one. The typical dynamic scheduler is one that learns about changes in a source repository (generally because some developer checks in some change), and triggers one or more builders in response to those changes. Let's assume for now that the scheduler "magically" learns about changes in the repository (more about this later); here's how we would define it:

from buildbot.plugins import schedulers

# define the dynamic scheduler
trunkchanged = schedulers.SingleBranchScheduler(name="trunkchanged",

# define the available schedulers
c['schedulers'] = [trunkchanged]

This scheduler receives changes happening to the repository, and among all of them, pays attention to those happening in "trunk" (that's what branch=None means). In other words, it filters the changes to react only to those it's interested in. When such changes are detected, and the tree has been quiet for 5 minutes (300 seconds), it runs the simplebuild builder. The treeStableTimer helps in those situations where commits tend to happen in bursts, which would otherwise result in multiple build requests queuing up.

What if we want to act on two branches (say, trunk and 7.2)? First we create two builders, one for each branch (see the builders paragraph above), then we create two dynamic schedulers:

from buildbot.plugins import schedulers

# define the dynamic scheduler for trunk
trunkchanged = schedulers.SingleBranchScheduler(name="trunkchanged",

# define the dynamic scheduler for the 7.2 branch
branch72changed = schedulers.SingleBranchScheduler(name="branch72changed",

# define the available schedulers
c['schedulers'] = [trunkchanged, branch72changed]

The syntax of the change filter is VCS-dependent (above is for SVN), but again once the idea is clear, the documentation has all the details. Another feature of the scheduler is that is can be told which changes, within those it's paying attention to, are important and which are not. For example, there may be a documentation directory in the branch the scheduler is watching, but changes under that directory should not trigger a build of the binary. This finer filtering is implemented by means of the fileIsImportant argument to the scheduler (full details in the docs and - alas - in the sources). Change sources

Earlier we said that a dynamic scheduler "magically" learns about changes; the final piece of the puzzle are change sources, which are precisely the elements in Buildbot whose task is to detect changes in the repository and communicate them to the schedulers. Note that periodic schedulers don't need a change source, since they only depend on elapsed time; dynamic schedulers, on the other hand, do need a change source.

A change source is generally configured with information about a source repository (which is where changes happen); a change source can watch changes at different levels in the hierarchy of the repository, so for example it is possible to watch the whole repository or a subset of it, or just a single branch. This determines the extent of the information that is passed down to the schedulers.

There are many ways a change source can learn about changes; it can periodically poll the repository for changes, or the VCS can be configured (for example through hook scripts triggered by commits) to push changes into the change source. While these two methods are probably the most common, they are not the only possibilities; it is possible for example to have a change source detect changes by parsing some email sent to a mailing list when a commit happens, and yet other methods exist. The manual again has the details.

To complete our example, here's a change source that polls a SVN repository every 2 minutes:

from buildbot.plugins import changes, util

svnpoller = changes.SVNPoller(repourl="svn://myrepo/projects/coolproject",

c['change_source'] = svnpoller

This poller watches the whole "coolproject" section of the repository, so it will detect changes in all the branches. We could have said:

repourl = "svn://myrepo/projects/coolproject/trunk"


repourl = "svn://myrepo/projects/coolproject/branches/7.2"

to watch only a specific branch.

To watch another project, you need to create another change source -- and you need to filter changes by project. For instance, when you add a change source watching project 'superproject' to the above example, you need to change:

trunkchanged = schedulers.SingleBranchScheduler(name="trunkchanged",
                                                # ...

to e.g.:

trunkchanged = schedulers.SingleBranchScheduler(name="trunkchanged",
                                                change_filter=filter.ChangeFilter(project="coolproject", branch=None),
                                                # ...

else coolproject will be built when there's a change in superproject.

Since we're watching more than one branch, we need a method to tell in which branch the change occurred when we detect one. This is what the split_file argument does, it takes a callable that Buildbot will call to do the job. The split_file_branches function, which comes with Buildbot, is designed for exactly this purpose so that's what the example above uses.

And of course this is all SVN-specific, but there are pollers for all the popular VCSs.

But note: if you have many projects, branches, and builders it probably pays to not hardcode all the schedulers and builders in the configuration, but generate them dynamically starting from list of all projects, branches, targets etc. and using loops to generate all possible combinations (or only the needed ones, depending on the specific setup), as explained in the documentation chapter about Customization. Status targets

Now that the basics are in place, let's go back to the builders, which is where the real work happens. Status targets are simply the means Buildbot uses to inform the world about what's happening, that is, how builders are doing. There are many status targets: a web interface, a mail notifier, an IRC notifier, and others. They are described fairly well in the manual.

One thing I've found useful is the ability to pass a domain name as the lookup argument to a mailNotifier, which allows you to take an unqualified username as it appears in the SVN change and create a valid email address by appending the given domain name to it:

from buildbot.plugins import status

# if jsmith commits a change, mail for the build is sent to
notifier = status.MailNotifier(fromaddr="",

The mail notifier can be customized at will by means of the messageFormatter argument, which is a function that Buildbot calls to format the body of the email, and to which it makes available lots of information about the build. Here all the details. Conclusion

Please note that this article has just scratched the surface; given the complexity of the task of build automation, the possibilities are almost endless. So there's much, much more to say about Buildbot. However, hopefully this is a preparation step before reading the official manual. Had I found an explanation as the one above when I was approaching Buildbot, I'd have had to read the manual just once, rather than multiple times. Hope this can help someone else.

(Thanks to Davide Brini for permission to include this tutorial, derived from one he originally posted at .)

This is the Buildbot manual for Buildbot version |version|.

2. Buildbot Manual

2.1. Introduction

Buildbot is a system to automate the compile/test cycle required by most software projects to validate code changes. By automatically rebuilding and testing the tree each time something has changed, build problems are pinpointed quickly, before other developers are inconvenienced by the failure. The guilty developer can be identified and harassed without human intervention. By running the builds on a variety of platforms, developers who do not have the facilities to test their changes everywhere before checkin will at least know shortly afterwards whether they have broken the build or not. Warning counts, lint checks, image size, compile time, and other build parameters can be tracked over time, are more visible, and are therefore easier to improve.

The overall goal is to reduce tree breakage and provide a platform to run tests or code-quality checks that are too annoying or pedantic for any human to waste their time with. Developers get immediate (and potentially public) feedback about their changes, encouraging them to be more careful about testing before checkin.


  • run builds on a variety of worker platforms
  • arbitrary build process: handles projects using C, Python, whatever
  • minimal host requirements: Python and Twisted
  • workers can be behind a firewall if they can still do checkout
  • status delivery through web page, email, IRC, other protocols
  • track builds in progress, provide estimated completion time
  • flexible configuration by subclassing generic build process classes
  • debug tools to force a new build, submit fake Changes, query worker status
  • released under the GPL

2.1.1. History and Philosophy

The Buildbot was inspired by a similar project built for a development team writing a cross-platform embedded system. The various components of the project were supposed to compile and run on several flavors of unix (linux, solaris, BSD), but individual developers had their own preferences and tended to stick to a single platform. From time to time, incompatibilities would sneak in (some unix platforms want to use string.h, some prefer strings.h), and then the tree would compile for some developers but not others. The buildbot was written to automate the human process of walking into the office, updating a tree, compiling (and discovering the breakage), finding the developer at fault, and complaining to them about the problem they had introduced. With multiple platforms it was difficult for developers to do the right thing (compile their potential change on all platforms); the buildbot offered a way to help.

Another problem was when programmers would change the behavior of a library without warning its users, or change internal aspects that other code was (unfortunately) depending upon. Adding unit tests to the codebase helps here: if an application's unit tests pass despite changes in the libraries it uses, you can have more confidence that the library changes haven't broken anything. Many developers complained that the unit tests were inconvenient or took too long to run: having the buildbot run them reduces the developer's workload to a minimum.

In general, having more visibility into the project is always good, and automation makes it easier for developers to do the right thing. When everyone can see the status of the project, developers are encouraged to keep the tree in good working order. Unit tests that aren't run on a regular basis tend to suffer from bitrot just like code does: exercising them on a regular basis helps to keep them functioning and useful.

The current version of the Buildbot is additionally targeted at distributed free-software projects, where resources and platforms are only available when provided by interested volunteers. The workers are designed to require an absolute minimum of configuration, reducing the effort a potential volunteer needs to expend to be able to contribute a new test environment to the project. The goal is for anyone who wishes that a given project would run on their favorite platform should be able to offer that project a worker, running on that platform, where they can verify that their portability code works, and keeps working.

2.1.2. System Architecture

The Buildbot consists of a single buildmaster and one or more workers, connected in a star topology. The buildmaster makes all decisions about what, when, and how to build. It sends commands to be run on the workers, which simply execute the commands and return the results. (certain steps involve more local decision making, where the overhead of sending a lot of commands back and forth would be inappropriate, but in general the buildmaster is responsible for everything).

The buildmaster is usually fed Changes by some sort of version control system (Change Sources), which may cause builds to be run. As the builds are performed, various status messages are produced, which are then sent to any registered Reporters.

Overview Diagram

The buildmaster is configured and maintained by the buildmaster admin, who is generally the project team member responsible for build process issues. Each worker is maintained by a worker admin, who do not need to be quite as involved. Generally workers are run by anyone who has an interest in seeing the project work well on their favorite platform. Worker Connections

The workers are typically run on a variety of separate machines, at least one per platform of interest. These machines connect to the buildmaster over a TCP connection to a publically-visible port. As a result, the workers can live behind a NAT box or similar firewalls, as long as they can get to buildmaster. The TCP connections are initiated by the worker and accepted by the buildmaster, but commands and results travel both ways within this connection. The buildmaster is always in charge, so all commands travel exclusively from the buildmaster to the worker.

To perform builds, the workers must typically obtain source code from a CVS/SVN/etc repository. Therefore they must also be able to reach the repository. The buildmaster provides instructions for performing builds, but does not provide the source code itself.

Worker Connections Buildmaster Architecture

The buildmaster consists of several pieces:

Buildmaster Architecture
Change Sources
Which create a Change object each time something is modified in the VC repository. Most ChangeSources listen for messages from a hook script of some sort. Some sources actively poll the repository on a regular basis. All Changes are fed to the schedulers.
Which decide when builds should be performed. They collect Changes into BuildRequests, which are then queued for delivery to Builders until a worker is available.
Which control exactly how each build is performed (with a series of BuildSteps, configured in a BuildFactory). Each Build is run on a single worker.
Status plugins
Which deliver information about the build results through protocols like HTTP, mail, and IRC.

Each Builder is configured with a list of Workers that it will use for its builds. These workers are expected to behave identically: the only reason to use multiple Workers for a single Builder is to provide a measure of load-balancing.

Within a single Worker, each Builder creates its own WorkerForBuilder instance. These WorkerForBuilders operate independently from each other. Each gets its own base directory to work in. It is quite common to have many Builders sharing the same worker. For example, there might be two workers: one for i386, and a second for PowerPC. There may then be a pair of Builders that do a full compile/test run, one for each architecture, and a lone Builder that creates snapshot source tarballs if the full builders complete successfully. The full builders would each run on a single worker, whereas the tarball creation step might run on either worker (since the platform doesn't matter when creating source tarballs). In this case, the mapping would look like:

Builder(full-i386)  ->  Workers(worker-i386)
Builder(full-ppc)   ->  Workers(worker-ppc)
Builder(source-tarball) -> Workers(worker-i386, worker-ppc)

and each Worker would have two WorkerForBuilders inside it, one for a full builder, and a second for the source-tarball builder.

Once a WorkerForBuilder is available, the Builder pulls one or more BuildRequests off its incoming queue. (It may pull more than one if it determines that it can merge the requests together; for example, there may be multiple requests to build the current HEAD revision). These requests are merged into a single Build instance, which includes the SourceStamp that describes what exact version of the source code should be used for the build. The Build is then randomly assigned to a free WorkerForBuilder and the build begins.

The behaviour when BuildRequests are merged can be customized, Collapsing Build Requests. Status Delivery Architecture

The buildmaster maintains a central Status object, to which various status plugins are connected. Through this Status object, a full hierarchy of build status objects can be obtained.

Status Delivery

The configuration file controls which status plugins are active. Each status plugin gets a reference to the top-level Status object. From there they can request information on each Builder, Build, Step, and LogFile. This query-on-demand interface is used by the html.Waterfall plugin to create the main status page each time a web browser hits the main URL.

The status plugins can also subscribe to hear about new Builds as they occur: this is used by the MailNotifier to create new email messages for each recently-completed Build.

The Status object records the status of old builds on disk in the buildmaster's base directory. This allows it to return information about historical builds.

There are also status objects that correspond to Schedulers and Workers. These allow status plugins to report information about upcoming builds, and the online/offline status of each worker.

2.1.3. Control Flow

A day in the life of the buildbot:

  • A developer commits some source code changes to the repository. A hook script or commit trigger of some sort sends information about this change to the buildmaster through one of its configured Change Sources. This notification might arrive via email, or over a network connection (either initiated by the buildmaster as it subscribes to changes, or by the commit trigger as it pushes Changes towards the buildmaster). The Change contains information about who made the change, what files were modified, which revision contains the change, and any checkin comments.
  • The buildmaster distributes this change to all of its configured schedulers. Any important changes cause the tree-stable-timer to be started, and the Change is added to a list of those that will go into a new Build. When the timer expires, a Build is started on each of a set of configured Builders, all compiling/testing the same source code. Unless configured otherwise, all Builds run in parallel on the various workers.
  • The Build consists of a series of Steps. Each Step causes some number of commands to be invoked on the remote worker associated with that Builder. The first step is almost always to perform a checkout of the appropriate revision from the same VC system that produced the Change. The rest generally perform a compile and run unit tests. As each Step runs, the worker reports back command output and return status to the buildmaster.
  • As the Build runs, status messages like "Build Started", "Step Started", "Build Finished", etc, are published to a collection of Status Targets. One of these targets is usually the HTML Waterfall display, which shows a chronological list of events, and summarizes the results of the most recent build at the top of each column. Developers can periodically check this page to see how their changes have fared. If they see red, they know that they've made a mistake and need to fix it. If they see green, they know that they've done their duty and don't need to worry about their change breaking anything.
  • If a MailNotifier status target is active, the completion of a build will cause email to be sent to any developers whose Changes were incorporated into this Build. The MailNotifier can be configured to only send mail upon failing builds, or for builds which have just transitioned from passing to failing. Other status targets can provide similar real-time notification via different communication channels, like IRC.

2.2. Installation

2.2.1. Buildbot Components

Buildbot is shipped in two components: the buildmaster (called buildbot for legacy reasons) and the worker. The worker component has far fewer requirements, and is more broadly compatible than the buildmaster. You will need to carefully pick the environment in which to run your buildmaster, but the worker should be able to run just about anywhere.

It is possible to install the buildmaster and worker on the same system, although for anything but the smallest installation this arrangement will not be very efficient.

2.2.2. Requirements Common Requirements

At a bare minimum, you'll need the following for both the buildmaster and a worker:


Both Buildbot master and Buildbot worker require Python-2.6, although Python-2.7 is recommended.


This should be a "normal" build of Python. Builds of Python with debugging enabled or other unusual build parameters are likely to cause incorrect behavior.


Buildbot requires Twisted-11.0.0 or later on the master, and Twisted-8.1.0 on the worker. In upcoming versions of Buildbot, a newer Twisted will also be required on the worker. As always, the most recent version is recommended. Note that Twisted requires ZopeInterface to be installed as well.


As part of ongoing (but as-yet incomplete) work to make Buildbot compatible with Python 3, the master requires the future module.

Of course, your project's build process will impose additional requirements on the workers. These hosts must have all the tools necessary to compile and test your project's source code.

Windows Support

Buildbot - both master and worker - runs well natively on Windows. The worker runs well on Cygwin, but because of problems with SQLite on Cygwin, the master does not.

Buildbot's windows testing is limited to the most recent Twisted and Python versions. For best results, use the most recent available versions of these libraries on Windows.


Twisted requires PyWin32 in order to spawn processes on Windows. Buildmaster Requirements

Note that all of these requirements aside from SQLite can easily be installed from the Python package repository, PyPI.


Buildbot requires a database to store its state, and by default uses SQLite. Version 3.7.0 or higher is recommended, although Buildbot will run down to 3.6.16 -- at the risk of "Database is locked" errors. The minimum version is 3.4.0, below which parallel database queries and schema introspection fail.

Please note that Python ships with sqlite3 by default since Python 2.6. Python2.6 for Windows ships with sqlite 3.6.2, thus you will not be able to run buildbot with sqlite on Windows and Python 2.6.

If you configure a different database engine, then SQLite is not required. however note that Buildbot's own unit tests require SQLite.


Buildbot requires Jinja version 2.1 or higher.

Jinja2 is a general purpose templating language and is used by Buildbot to generate the HTML output.


Buildbot requires SQLAlchemy version 0.8.0 or higher. SQLAlchemy allows Buildbot to build database schemas and queries for a wide variety of database systems.


Buildbot requires SQLAlchemy-Migrate version 0.9.0 or higher. Buildbot uses SQLAlchemy-Migrate to manage schema upgrades from version to version.


Buildbot requires Python-Dateutil in version 1.5 or higher (the last version to support Python-2.x). This is a small, pure-Python library.


The master requires Autobahn version 0.10.2 or higher

2.2.3. Installing the code The Buildbot Packages

Buildbot comes in several parts: buildbot (the buildmaster), buildbot-worker (the worker), buildbot-www, and several web plugins such as buildbot-waterfall-view.

The worker and buildmaster can be installed individually or together. The base web (buildbot.www) and web plugins are required to run a master with a web interface (the common configuration). Installation From PyPI

The preferred way to install Buildbot is using pip. For the master:

pip install buildbot

and for the worker:

pip install buildbot-worker

When using pip to install instead of distribution specific package manangers, e.g. via apt-get or ports, it is simpler to choose exactly which version one wants to use. It may however be easier to install via distribution specific package mangers but note that they may provide an earlier version than what is available via pip.

If you plan to use TLS or SSL in master configuration (e.g. to fetch resources over HTTPS using twisted.web.client), you need to install Buildbot with tls extras:

pip install buildbot[tls] Installation From Tarballs

Buildbot master and buildbot-worker are installed using the standard Python distutils process. For either component, after unpacking the tarball, the process is:

python build
python install

where the install step may need to be done as root. This will put the bulk of the code in somewhere like /usr/lib/pythonx.y/site-packages/buildbot. It will also install the buildbot command-line tool in /usr/bin/buildbot.

If the environment variable $NO_INSTALL_REQS is set to 1, then will not try to install Buildbot's requirements. This is usually only useful when building a Buildbot package.

To test this, shift to a different directory (like /tmp), and run:

buildbot --version
# or
buildbot-worker --version

If it shows you the versions of Buildbot and Twisted, the install went ok. If it says "no such command" or it gets an ImportError when it tries to load the libraries, then something went wrong. pydoc buildbot is another useful diagnostic tool.

Windows users will find these files in other places. You will need to make sure that Python can find the libraries, and will probably find it convenient to have buildbot on your PATH. Installation in a Virtualenv

If you cannot or do not wish to install the buildbot into a site-wide location like /usr or /usr/local, you can also install it into the account's home directory or any other location using a tool like virtualenv. Running Buildbot's Tests (optional)

If you wish, you can run the buildbot unit test suite. First, ensure you have the mock Python module installed from PyPI. You must not be using a Python wheels packaged version of Buildbot or have specified the bdist_wheel command when building. The test suite is not included with the PyPi packaged version. This module is not required for ordinary Buildbot operation - only to run the tests. Note that this is not the same as the Fedora mock package!

You can check with

python -mmock

Then, run the tests:

PYTHONPATH=. trial buildbot.test
# or
PYTHONPATH=. trial buildbot_worker.test

Nothing should fail, although a few might be skipped.

If any of the tests fail for reasons other than a missing mock, you should stop and investigate the cause before continuing the installation process, as it will probably be easier to track down the bug early. In most cases, the problem is incorrectly installed Python modules or a badly configured PYTHONPATH. This may be a good time to contact the Buildbot developers for help.

2.2.4. Upgrading to Nine

Upgrading a Buildbot instance from 0.8.x to 0.9.x may require a number of changes to the master configuration. Those changes are summarized here. If you are starting fresh with 0.9.0 or later, you can safely skip this section. Config File Syntax

In preparation for compatibility with Python 3, Buildbot coniguration files no longer allow the print statement:

print "foo"

To fix, simply enclose the print arguments in parentheses:

print("foo") Plugins

Although plugin support was available in 0.8.12, its use is now highly recommended. Instead of importing modules directly in master.cfg, import the plugin kind from buildbot.plugins:

from buildbot.plugins import steps

Then access the plugin itself as an attribute:


See Plugin Infrastructure in Buildbot for more information. Web Status

The most prominent change is that the existing WebStatus class is now gone, replaced by the new www functionality.

Thus an html.WebStatus entry in c['status'] should be removed and replaced with configuration in c['www']`. For example, replace:

from buildbot.status import html
c['status'].append(html.WebStatus(http_port=8010, allowForce=True)


c['www'] = dict(port=8010,

See www for more information. Status Classes

Where in 0.8.x most of the data about a build was available synchronously, it must now be fetched dynamically using the Data API. All classes under the Python package buildbot.status should be considered deprecated. Many have already been removed, and the remainder have limited functionality. Any custom code which refers to these classes must be rewritten to use the Data API. Avoid the temptatation to reach into the Buildbot source code to find other useful-looking methods!

Common uses of the status API are:

  • getBuild in a custom renderable
  • MailNotifier message formatters (see below for upgrade hints)
  • doIf funtions on steps

Import paths for several classes under the buildbot.status package but which remain useful have changed. Most of these are now available as plugins (see above), but for the remainder, consult the source code. BuildRequest Merging

Buildbot 0.9.x has replaced the old concept of request merging (mergeRequests) with a more flexible request-collapsing mechanism. See collapseRequests for more information. Status Reporters

In fact, the whole c['status'] configuration parameter is gone.

Many of the status listeners used in the status hierarchy in 0.8.x have been replaced with "reporters" that are availabale as buildbot plugins. However, note that not all status listeners have yet been ported. See the release notes for details.

Including the "status" key in the configuration object will cause a configuration error. All reporters should be included in c['services'] as described in Reporters.

The available reporters as of 0.9.0 are

See the reporter index for the full, current list.

A few notes on changes to the configuration of these reporters:

  • MailNotifier argument messageFormatter should now be a buildbot.reporters.message.MessageFormatter, due to the removal of the status classes (see above), such formatters must be re-implemented using the Data API.
  • MailNotifier argument previousBuildGetter is not supported anymore
  • MailNotifier no longer forces SSL 3.0 when useTls is true.
  • GerritStatusPush callbacks slightly changed signature, and include a master reference instead of a status reference.
  • GitHubStatusPush now accepts a context parameter to be passed to the GitHub Status API.
  • buildbot.status.builder.Results and the constants buildbot.status.results.SUCCESS should be imported from the buildbot.process.results module instead. Steps

Buildbot-0.8.9 introduced "new-style steps", with an asynchronous run method. In the remaining 0.8.x releases, use of new-style and old-style steps were supported side-by-side. In 0.9.x, old-style steps are emulated using a collection of hacks to allow asynchronous calls to be called from synchronous code. This emulation is imperfect, and you are strongly encouraged to rewrite any custom steps as New-Style Build Steps.

Note that new-style steps now "push" their status when it changes, so the describe method no longer exists. Identifiers

Many strings in Buildbot must now be identifiers. Identifiers are designed to fit easily and unambiguously into URLs, AMQP routes, and the like. An "identifier" is a nonempty unicode string of limited length, containing only ASCII alphanumeric characters along with - (dash) and _ (underscore), and not beginning with a digit

Unfortunately, many existing names do not fit this pattern.

The following fields are identifiers:

  • worker name (50-character)
  • builder name (20-character)
  • step name (50-character) Serving static files

Since version 0.9.0 Buildbot doesn't use and don't serve master's public_html directory. You need to use third-party HTTP server for serving static files. Transition to "worker" terminology

Since version 0.9.0 of Buildbot "slave"-based terminology is deprecated in favor of "worker"-based terminology.

All identifiers, messages and documentation were updated to use "worker" instead of "slave". Old API names are still available, but deprecated.

For details about changed API and how to control generated warnings see Transition to "worker" terminology. Other Config Settings

The default master.cfg file contains some new changes, which you should look over:

  • c['protocols'] = {'pb': {'port': 9989}} (the default port used by the workers)
  • Waterfall View: requires installation (pip install buildbot-waterfall-view) and configuration (c['www'] = { ..., 'plugins': {'waterfall_view': {} }). Build History

There is no support for importing build history from 0.8.x (where the history was stored on-disk in pickle files) into 0.9.x (where it is stored in the database). More Information

For minor changes not mentioned here, consult the release notes for the versions over which you are upgrading.

Buildbot-0.9.0 represents several years' work, and as such we may have missed potential migration issues. To find the latest "gotchas" and share with other users, see

2.2.5. Buildmaster Setup Creating a buildmaster

As you learned earlier (System Architecture), the buildmaster runs on a central host (usually one that is publicly visible, so everybody can check on the status of the project), and controls all aspects of the buildbot system

You will probably wish to create a separate user account for the buildmaster, perhaps named buildmaster. Do not run the buildmaster as root!

You need to choose a directory for the buildmaster, called the basedir. This directory will be owned by the buildmaster. It will contain configuration, the database, and status information - including logfiles. On a large buildmaster this directory will see a lot of activity, so it should be on a disk with adequate space and speed.

Once you've picked a directory, use the buildbot create-master command to create the directory and populate it with startup files:

buildbot create-master -r basedir

You will need to create a configuration file before starting the buildmaster. Most of the rest of this manual is dedicated to explaining how to do this. A sample configuration file is placed in the working directory, named master.cfg.sample, which can be copied to master.cfg and edited to suit your purposes.

(Internal details: This command creates a file named buildbot.tac that contains all the state necessary to create the buildmaster. Twisted has a tool called twistd which can use this .tac file to create and launch a buildmaster instance. Twistd takes care of logging and daemonization (running the program in the background). /usr/bin/buildbot is a front end which runs twistd for you.)

Your master will need a database to store the various information about your builds, and its configuration. By default, the sqlite3 backend will be used. This needs no configuration, neither extra software. All information will be stored in the file state.sqlite. Buildbot however supports multiple backends. See Using A Database Server for more options.

Buildmaster Options

This section lists options to the create-master command. You can also type buildbot create-master --help for an up-to-the-moment summary.


This option will allow to re-use an existing directory.


This disables internal worker log management mechanism. With this option worker does not override the default logfile name and its behaviour giving a possibility to control those with command-line options of twistd daemon.


This creates a "relocatable" buildbot.tac, which uses relative paths instead of absolute paths, so that the buildmaster directory can be moved about.


The name of the configuration file to use. This configuration file need not reside in the buildmaster directory.


This is the size in bytes when to rotate the Twisted log files. The default is 10MiB.


This is the number of log rotations to keep around. You can either specify a number or None to keep all twistd.log files around. The default is 10.


The database that the Buildmaster should use. Note that the same value must be added to the configuration file. Upgrading an Existing Buildmaster

If you have just installed a new version of the Buildbot code, and you have buildmasters that were created using an older version, you'll need to upgrade these buildmasters before you can use them. The upgrade process adds and modifies files in the buildmaster's base directory to make it compatible with the new code.

buildbot upgrade-master basedir

This command will also scan your master.cfg file for incompatibilities (by loading it and printing any errors or deprecation warnings that occur). Each buildbot release tries to be compatible with configurations that worked cleanly (i.e. without deprecation warnings) on the previous release: any functions or classes that are to be removed will first be deprecated in a release, to give you a chance to start using the replacement.

The upgrade-master command is idempotent. It is safe to run it multiple times. After each upgrade of the buildbot code, you should use upgrade-master on all your buildmasters.

In general, Buildbot workers and masters can be upgraded independently, although some new features will not be available, depending on the master and worker versions.

Beyond this general information, read all of the sections below that apply to versions through which you are upgrading.

Version-specific Notes
Upgrading from Buildbot-0.8.x to Buildbot-0.9.x

See Upgrading to Nine for a guide to upgrading from 0.8.x to 0.9.x

Upgrading a Buildmaster to Buildbot-0.7.6

The 0.7.6 release introduced the public_html/ directory, which contains index.html and other files served by the WebStatus and Waterfall status displays. The upgrade-master command will create these files if they do not already exist. It will not modify existing copies, but it will write a new copy in e.g. if the new version differs from the version that already exists.

Upgrading a Buildmaster to Buildbot-0.8.0

Buildbot-0.8.0 introduces a database backend, which is SQLite by default. The upgrade-master command will automatically create and populate this database with the changes the buildmaster has seen. Note that, as of this release, build history is not contained in the database, and is thus not migrated.

Upgrading into a non-SQLite database

If you are not using sqlite, you will need to add an entry into your master.cfg to reflect the database version you are using. The upgrade process does not edit your master.cfg for you. So something like:

# for using mysql:
c['db_url'] = 'mysql://bbuser:<password>@localhost/buildbot'

Once the parameter has been added, invoke upgrade-master. This will extract the DB url from your configuration file.

buildbot upgrade-master

See Database Specification for more options to specify a database.

2.2.6. Worker Setup Creating a worker

Typically, you will be adding a worker to an existing buildmaster, to provide additional architecture coverage. The buildbot administrator will give you several pieces of information necessary to connect to the buildmaster. You should also be somewhat familiar with the project being tested, so you can troubleshoot build problems locally.

The buildbot exists to make sure that the project's stated how to build it process actually works. To this end, the worker should run in an environment just like that of your regular developers. Typically the project build process is documented somewhere (README, INSTALL, etc), in a document that should mention all library dependencies and contain a basic set of build instructions. This document will be useful as you configure the host and account in which the worker runs.

Here's a good checklist for setting up a worker:

  1. Set up the account
It is recommended (although not mandatory) to set up a separate user account for the worker. This account is frequently named buildbot or worker. This serves to isolate your personal working environment from that of the worker's, and helps to minimize the security threat posed by letting possibly-unknown contributors run arbitrary code on your system. The account should have a minimum of fancy init scripts.
  1. Install the buildbot code
Follow the instructions given earlier (Installing the code). If you use a separate worker account, and you didn't install the buildbot code to a shared location, then you will need to install it with --home=~ for each account that needs it.
  1. Set up the host

Make sure the host can actually reach the buildmaster. Usually the buildmaster is running a status webserver on the same machine, so simply point your web browser at it and see if you can get there. Install whatever additional packages or libraries the project's INSTALL document advises. (or not: if your worker is supposed to make sure that building without optional libraries still works, then don't install those libraries.)

Again, these libraries don't necessarily have to be installed to a site-wide shared location, but they must be available to your build process. Accomplishing this is usually very specific to the build process, so installing them to /usr or /usr/local is usually the best approach.

  1. Test the build process
Follow the instructions in the INSTALL document, in the worker's account. Perform a full CVS (or whatever) checkout, configure, make, run tests, etc. Confirm that the build works without manual fussing. If it doesn't work when you do it by hand, it will be unlikely to work when the buildbot attempts to do it in an automated fashion.
  1. Choose a base directory
This should be somewhere in the worker's account, typically named after the project which is being tested. The worker will not touch any file outside of this directory. Something like ~/Buildbot or ~/Workers/fooproject is appropriate.
  1. Get the buildmaster host/port, botname, and password

When the buildbot admin configures the buildmaster to accept and use your worker, they will provide you with the following pieces of information:

  1. Create the worker

Now run the 'worker' command as follows:


This will create the base directory and a collection of files inside, including the buildbot.tac file that contains all the information you passed to the buildbot command.

  1. Fill in the hostinfo files

When it first connects, the worker will send a few files up to the buildmaster which describe the host that it is running on. These files are presented on the web status display so that developers have more information to reproduce any test failures that are witnessed by the buildbot. There are sample files in the info subdirectory of the buildbot's base directory. You should edit these to correctly describe you and your host.

BASEDIR/info/admin should contain your name and email address. This is the worker admin address, and will be visible from the build status page (so you may wish to munge it a bit if address-harvesting spambots are a concern).

BASEDIR/info/host should be filled with a brief description of the host: OS, version, memory size, CPU speed, versions of relevant libraries installed, and finally the version of the buildbot code which is running the worker.

The optional BASEDIR/info/access_uri can specify a URI which will connect a user to the machine. Many systems accept ssh://hostname URIs for this purpose.

If you run many workers, you may want to create a single ~worker/info file and share it among all the workers with symlinks.

Worker Options

There are a handful of options you might want to use when creating the worker with the buildbot-worker create-worker <options> DIR <params> command. You can type buildbot-worker create-worker --help for a summary. To use these, just include them on the buildbot-worker create-worker command line, like this

buildbot-worker create-worker --umask=022 ~/worker {myworkername} {mypasswd}

This disables internal worker log management mechanism. With this option worker does not override the default logfile name and its behaviour giving a possibility to control those with command-line options of twistd daemon.


This is a string (generally an octal representation of an integer) which will cause the worker process' umask value to be set shortly after initialization. The twistd daemonization utility forces the umask to 077 at startup (which means that all files created by the worker or its child processes will be unreadable by any user other than the worker account). If you want build products to be readable by other accounts, you can add --umask=022 to tell the worker to fix the umask after twistd clobbers it. If you want build products to be writable by other accounts too, use --umask=000, but this is likely to be a security problem.


This is a number that indicates how frequently keepalive messages should be sent from the worker to the buildmaster, expressed in seconds. The default (600) causes a message to be sent to the buildmaster at least once every 10 minutes. To set this to a lower value, use e.g. --keepalive=120.

If the worker is behind a NAT box or stateful firewall, these messages may help to keep the connection alive: some NAT boxes tend to forget about a connection if it has not been used in a while. When this happens, the buildmaster will think that the worker has disappeared, and builds will time out. Meanwhile the worker will not realize than anything is wrong.


This is a number that indicates the maximum amount of time the worker will wait between connection attempts, expressed in seconds. The default (300) causes the worker to wait at most 5 minutes before trying to connect to the buildmaster again.


This is the size in bytes when to rotate the Twisted log files.


This is the number of log rotations to keep around. You can either specify a number or None to keep all twistd.log files around. The default is 10.


Can also be passed directly to the Worker constructor in buildbot.tac. If set, it allows the worker to initiate a graceful shutdown, meaning that it will ask the master to shut down the worker when the current build, if any, is complete.

Setting allow_shutdown to file will cause the worker to watch shutdown.stamp in basedir for updates to its mtime. When the mtime changes, the worker will request a graceful shutdown from the master. The file does not need to exist prior to starting the worker.

Setting allow_shutdown to signal will set up a SIGHUP handler to start a graceful shutdown. When the signal is received, the worker will request a graceful shutdown from the master.

The default value is None, in which case this feature will be disabled.

Both master and worker must be at least version 0.8.3 for this feature to work.

Other Worker Configuration

This represents the encoding that buildbot should use when converting unicode commandline arguments into byte strings in order to pass to the operating system when spawning new processes.

The default value is what Python's sys.getfilesystemencoding returns, which on Windows is 'mbcs', on Mac OSX is 'utf-8', and on Unix depends on your locale settings.

If you need a different encoding, this can be changed in your worker's buildbot.tac file by adding a unicode_encoding argument to the Worker constructor.

s = Worker(buildmaster_host, port, workername, passwd, basedir,
           keepalive, usepty, umask=umask, maxdelay=maxdelay,
           unicode_encoding='utf-8', allow_shutdown='signal') Upgrading an Existing Worker
Version-specific Notes

During project lifetime worker has transitioned over few states:

  1. Before Buildbot version 0.8.1 worker were integral part of buildbot package distribution.
  2. Starting from Buildbot version 0.8.1 worker were extracted from buildbot package to buildbot-slave package.
  3. Starting from Buildbot version 0.9.0 the buildbot-slave package was renamed to buildbot-worker.
Upgrading a Worker to buildbot-slave 0.8.1

Before Buildbot version 0.8.1, the Buildbot master and worker were part of the same distribution. As of version 0.8.1, the worker is a separate distribution.

As of this release, you will need to install buildbot-slave to run a worker.

Any automatic startup scripts that had run buildbot start for previous versions should be changed to run buildslave start instead.

If you are running a version later than 0.8.1, then you can skip the remainder of this section: the upgrade-slave command will take care of this. If you are upgrading directly to 0.8.1, read on.

The existing buildbot.tac for any workers running older versions will need to be edited or replaced. If the loss of cached worker state (e.g., for Source steps in copy mode) is not problematic, the easiest solution is to simply delete the worker directory and re-run buildslave create-slave.

If deleting the worker directory is problematic, the change to buildbot.tac is simple. On line 3, replace:

from import BuildSlave


from import BuildSlave

After this change, the worker should start as usual.

Upgrading from 0.8.1 to the latest 0.8.* version of buildbot-slave

If you have just installed a new version of Buildbot-slave, you may need to take some steps to upgrade it. If you are upgrading to version 0.8.2 or later, you can run

buildslave upgrade-slave /path/to/worker/dir
Upgrading from the latest version of buildbot-slave to buildbot-worker

If the loss of cached worker state (e.g., for Source steps in copy mode) is not problematic, the easiest solution is to simply delete the worker directory and re-run buildbot-worker create-worker.

If deleting the worker directory is problematic, you can change buildbot.tac in the following way:

  1. Replace:

    from import BuildSlave


    from import Worker
  2. Replace:

    application = service.Application('buildslave')


    application = service.Application('buildbot-worker')
  3. Replace:

    s = BuildSlave(buildmaster_host, port, slavename, passwd, basedir,
                   keepalive, usepty, umask=umask, maxdelay=maxdelay,
                   numcpus=numcpus, allow_shutdown=allow_shutdown)


    s = Worker(buildmaster_host, port, slavename, passwd, basedir,
               keepalive, umask=umask, maxdelay=maxdelay,
               numcpus=numcpus, allow_shutdown=allow_shutdown)

See Transition to "Worker" Terminology for details of changes in version Buildbot 0.9.0.

2.2.7. Next Steps Launching the daemons

Both the buildmaster and the worker run as daemon programs. To launch them, pass the working directory to the buildbot and buildbot-worker commands, as appropriate:

# start a master
buildbot start [ BASEDIR ]
# start a worker
buildbot-worker start [ WORKER_BASEDIR ]

The BASEDIR is option and can be omitted if the current directory contains the buildbot configuration (the buildbot.tac file).

buildbot start

This command will start the daemon and then return, so normally it will not produce any output. To verify that the programs are indeed running, look for a pair of files named twistd.log and that should be created in the working directory. contains the process ID of the newly-spawned daemon.

When the worker connects to the buildmaster, new directories will start appearing in its base directory. The buildmaster tells the worker to create a directory for each Builder which will be using that worker. All build operations are performed within these directories: CVS checkouts, compiles, and tests.

Once you get everything running, you will want to arrange for the buildbot daemons to be started at boot time. One way is to use cron, by putting them in a @reboot crontab entry [1]

@reboot buildbot start [ BASEDIR ]

When you run crontab to set this up, remember to do it as the buildmaster or worker account! If you add this to your crontab when running as your regular account (or worse yet, root), then the daemon will run as the wrong user, quite possibly as one with more authority than you intended to provide.

It is important to remember that the environment provided to cron jobs and init scripts can be quite different that your normal runtime. There may be fewer environment variables specified, and the PATH may be shorter than usual. It is a good idea to test out this method of launching the worker by using a cron job with a time in the near future, with the same command, and then check twistd.log to make sure the worker actually started correctly. Common problems here are for /usr/local or ~/bin to not be on your PATH, or for PYTHONPATH to not be set correctly. Sometimes HOME is messed up too.

Some distributions may include conveniences to make starting buildbot at boot time easy. For instance, with the default buildbot package in Debian-based distributions, you may only need to modify /etc/default/buildbot (see also /etc/init.d/buildbot, which reads the configuration in /etc/default/buildbot).

Buildbot also comes with its own init scripts that provide support for controlling multi-worker and multi-master setups (mostly because they are based on the init script from the Debian package). With a little modification these scripts can be used both on Debian and RHEL-based distributions and may thus prove helpful to package maintainers who are working on buildbot (or those that haven't yet split buildbot into master and worker packages).

# install as /etc/default/buildbot-worker
#         or /etc/sysconfig/buildbot-worker

# install as /etc/default/buildmaster
#         or /etc/sysconfig/buildmaster

# install as /etc/init.d/buildbot-worker

# install as /etc/init.d/buildmaster

# ... and tell sysvinit about them
chkconfig buildmaster reset
# ... or
update-rc.d buildmaster defaults Logfiles

While a buildbot daemon runs, it emits text to a logfile, named twistd.log. A command like tail -f twistd.log is useful to watch the command output as it runs.

The buildmaster will announce any errors with its configuration file in the logfile, so it is a good idea to look at the log at startup time to check for any problems. Most buildmaster activities will cause lines to be added to the log. Shutdown

To stop a buildmaster or worker manually, use:

buildbot stop [ BASEDIR ]
# or
buildbot-worker stop [ WORKER_BASEDIR ]

This simply looks for the file and kills whatever process is identified within.

At system shutdown, all processes are sent a SIGKILL. The buildmaster and worker will respond to this by shutting down normally.

The buildmaster will respond to a SIGHUP by re-reading its config file. Of course, this only works on Unix-like systems with signal support, and won't work on Windows. The following shortcut is available:

buildbot reconfig [ BASEDIR ]

When you update the Buildbot code to a new release, you will need to restart the buildmaster and/or worker before it can take advantage of the new code. You can do a buildbot stop BASEDIR and buildbot start BASEDIR in quick succession, or you can use the restart shortcut, which does both steps for you:

buildbot restart [ BASEDIR ]

Workers can similarly be restarted with:

buildbot-worker restart [ BASEDIR ]

There are certain configuration changes that are not handled cleanly by buildbot reconfig. If this occurs, buildbot restart is a more robust tool to fully switch over to the new configuration.

buildbot restart may also be used to start a stopped Buildbot instance. This behaviour is useful when writing scripts that stop, start and restart Buildbot.

A worker may also be gracefully shutdown from the web UI. This is useful to shutdown a worker without interrupting any current builds. The buildmaster will wait until the worker is finished all its current builds, and will then tell the worker to shutdown.

[1]This @reboot syntax is understood by Vixie cron, which is the flavor usually provided with Linux systems. Other unices may have a cron that doesn't understand @reboot

2.3. Concepts

This chapter defines some of the basic concepts that the Buildbot uses. You'll need to understand how the Buildbot sees the world to configure it properly.

2.3.1. Source Stamps

Source code comes from repositories, provided by version control systems. Repositories are generally identified by URLs, e.g., git://

In these days of distributed version control systems, the same codebase may appear in multiple repositories. For example, and both contain the Firefox codebase, although not exactly the same code.

Many projects are built from multiple codebases. For example, a company may build several applications based on the same core library. The "app" codebase and the "core" codebase are in separate repositories, but are compiled together and constitute a single project. Changes to either codebase should cause a rebuild of the application.

Most version control systems define some sort of revision that can be used (sometimes in combination with a branch) to uniquely specify a particular version of the source code.

To build a project, Buildbot needs to know exactly which version of each codebase it should build. It uses a source stamp to do so for each codebase; the collection of sourcestamps required for a project is called a source stamp set.

2.3.2. Version Control Systems

Buildbot supports a significant number of version control systems, so it treats them abstractly.

For purposes of deciding when to perform builds, Buildbot's change sources monitor repositories, and represent any updates to those repositories as changes. These change sources fall broadly into two categories: pollers which periodically check the repository for updates; and hooks, where the repository is configured to notify Buildbot whenever an update occurs.

This concept does not map perfectly to every version control system. For example, for CVS Buildbot must guess that version updates made to multiple files within a short time represent a single change; for DVCS's like Git, Buildbot records a change when a commit is pushed to the monitored repository, not when it is initially committed. We assume that the Changes arrive at the master in the same order in which they are committed to the repository.

When it comes time to actually perform a build, a scheduler prepares a source stamp set, as described above, based on its configuration. When the build begins, one or more source steps use the information in the source stamp set to actually check out the source code, using the normal VCS commands. Tree Stability

Changes tend to arrive at a buildmaster in bursts. In many cases, these bursts of changes are meant to be taken together. For example, a developer may have pushed multiple commits to a DVCS that comprise the same new feature or bugfix. To avoid trying to build every change, Buildbot supports the notion of tree stability, by waiting for a burst of changes to finish before starting to schedule builds. This is implemented as a timer, with builds not scheduled until no changes have occurred for the duration of the timer. How Different VC Systems Specify Sources

For CVS, the static specifications are repository and module. In addition to those, each build uses a timestamp (or omits the timestamp to mean the latest) and branch tag (which defaults to HEAD). These parameters collectively specify a set of sources from which a build may be performed.

Subversion, combines the repository, module, and branch into a single Subversion URL parameter. Within that scope, source checkouts can be specified by a numeric revision number (a repository-wide monotonically-increasing marker, such that each transaction that changes the repository is indexed by a different revision number), or a revision timestamp. When branches are used, the repository and module form a static baseURL, while each build has a revision number and a branch (which defaults to a statically-specified defaultBranch). The baseURL and branch are simply concatenated together to derive the repourl to use for the checkout.

Perforce is similar. The server is specified through a P4PORT parameter. Module and branch are specified in a single depot path, and revisions are depot-wide. When branches are used, the p4base and defaultBranch are concatenated together to produce the depot path.

Bzr (which is a descendant of Arch/Bazaar, and is frequently referred to as "Bazaar") has the same sort of repository-vs-workspace model as Arch, but the repository data can either be stored inside the working directory or kept elsewhere (either on the same machine or on an entirely different machine). For the purposes of Buildbot (which never commits changes), the repository is specified with a URL and a revision number.

The most common way to obtain read-only access to a bzr tree is via HTTP, simply by making the repository visible through a web server like Apache. Bzr can also use FTP and SFTP servers, if the worker process has sufficient privileges to access them. Higher performance can be obtained by running a special Bazaar-specific server. None of these matter to the buildbot: the repository URL just has to match the kind of server being used. The repoURL argument provides the location of the repository.

Branches are expressed as subdirectories of the main central repository, which means that if branches are being used, the BZR step is given a baseURL and defaultBranch instead of getting the repoURL argument.

Darcs doesn't really have the notion of a single master repository. Nor does it really have branches. In Darcs, each working directory is also a repository, and there are operations to push and pull patches from one of these repositories to another. For the Buildbot's purposes, all you need to do is specify the URL of a repository that you want to build from. The worker will then pull the latest patches from that repository and build them. Multiple branches are implemented by using multiple repositories (possibly living on the same server).

Builders which use Darcs therefore have a static repourl which specifies the location of the repository. If branches are being used, the source Step is instead configured with a baseURL and a defaultBranch, and the two strings are simply concatenated together to obtain the repository's URL. Each build then has a specific branch which replaces defaultBranch, or just uses the default one. Instead of a revision number, each build can have a context, which is a string that records all the patches that are present in a given tree (this is the output of darcs changes --context, and is considerably less concise than, e.g. Subversion's revision number, but the patch-reordering flexibility of Darcs makes it impossible to provide a shorter useful specification).

Mercurial is like Darcs, in that each branch is stored in a separate repository. The repourl, baseURL, and defaultBranch arguments are all handled the same way as with Darcs. The revision, however, is the hash identifier returned by hg identify.

Git also follows a decentralized model, and each repository can have several branches and tags. The source Step is configured with a static repourl which specifies the location of the repository. In addition, an optional branch parameter can be specified to check out code from a specific branch instead of the default master branch. The revision is specified as a SHA1 hash as returned by e.g. git rev-parse. No attempt is made to ensure that the specified revision is actually a subset of the specified branch.

Monotone is another that follows a decentralized model where each repository can have several branches and tags. The source Step is configured with static repourl and branch parameters, which specifies the location of the repository and the branch to use. The revision is specified as a SHA1 hash as returned by e.g. mtn automate select w:. No attempt is made to ensure that the specified revision is actually a subset of the specified branch.

2.3.3. Changes Who

Each Change has a who attribute, which specifies which developer is responsible for the change. This is a string which comes from a namespace controlled by the VC repository. Frequently this means it is a username on the host which runs the repository, but not all VC systems require this. Each StatusNotifier will map the who attribute into something appropriate for their particular means of communication: an email address, an IRC handle, etc.

This who attribute is also parsed and stored into Buildbot's database (see User Objects). Currently, only who attributes in Changes from git repositories are translated into user objects, but in the future all incoming Changes will have their who parsed and stored. Files

It also has a list of files, which are just the tree-relative filenames of any files that were added, deleted, or modified for this Change. These filenames are used by the fileIsImportant function (in the scheduler) to decide whether it is worth triggering a new build or not, e.g. the function could use the following function to only run a build if a C file were checked in:

def has_C_files(change):
    for name in change.files:
        if name.endswith(".c"):
            return True
    return False

Certain BuildSteps can also use the list of changed files to run a more targeted series of tests, e.g. the python_twisted.Trial step can run just the unit tests that provide coverage for the modified .py files instead of running the full test suite. Comments

The Change also has a comments attribute, which is a string containing any checkin comments. Project

The project attribute of a change or source stamp describes the project to which it corresponds, as a short human-readable string. This is useful in cases where multiple independent projects are built on the same buildmaster. In such cases, it can be used to control which builds are scheduled for a given commit, and to limit status displays to only one project. Repository

This attribute specifies the repository in which this change occurred. In the case of DVCS's, this information may be required to check out the committed source code. However, using the repository from a change has security risks: if Buildbot is configured to blindly trust this information, then it may easily be tricked into building arbitrary source code, potentially compromising the workers and the integrity of subsequent builds. Codebase

This attribute specifies the codebase to which this change was made. As described above, multiple repositories may contain the same codebase. A change's codebase is usually determined by the codebaseGenerator configuration. By default the codebase is ''; this value is used automatically for single-codebase configurations. Revision

Each Change can have a revision attribute, which describes how to get a tree with a specific state: a tree which includes this Change (and all that came before it) but none that come after it. If this information is unavailable, the revision attribute will be None. These revisions are provided by the ChangeSource.

Revisions are always strings.

revision is the seconds since the epoch as an integer.
revision is the revision number
revision is a large string, the output of darcs changes --context
revision is a short string (a hash ID), the output of hg identify
revision is the transaction number
revision is a short string (a SHA1 hash), the output of e.g. git rev-parse Branches

The Change might also have a branch attribute. This indicates that all of the Change's files are in the same named branch. The schedulers get to decide whether the branch should be built or not.

For VC systems like CVS, Git and Monotone the branch name is unrelated to the filename. (That is, the branch name and the filename inhabit unrelated namespaces.) For SVN, branches are expressed as subdirectories of the repository, so the file's repourl is a combination of some base URL, the branch name, and the filename within the branch. (In a sense, the branch name and the filename inhabit the same namespace.) Darcs branches are subdirectories of a base URL just like SVN. Mercurial branches are the same as Darcs.

branch='warner-newfeature', files=['src/foo.c']
branch='branches/warner-newfeature', files=['src/foo.c']
branch='warner-newfeature', files=['src/foo.c']
branch='warner-newfeature', files=['src/foo.c']
branch='warner-newfeature', files=['src/foo.c']
branch='warner-newfeature', files=['src/foo.c'] Change Properties

A Change may have one or more properties attached to it, usually specified through the Force Build form or sendchange. Properties are discussed in detail in the Build Properties section.

2.3.4. Scheduling Builds

Each Buildmaster has a set of scheduler objects, each of which gets a copy of every incoming Change. The Schedulers are responsible for deciding when Builds should be run. Some Buildbot installations might have a single scheduler, while others may have several, each for a different purpose.

For example, a quick scheduler might exist to give immediate feedback to developers, hoping to catch obvious problems in the code that can be detected quickly. These typically do not run the full test suite, nor do they run on a wide variety of platforms. They also usually do a VC update rather than performing a brand-new checkout each time.

A separate full scheduler might run more comprehensive tests, to catch more subtle problems. configured to run after the quick scheduler, to give developers time to commit fixes to bugs caught by the quick scheduler before running the comprehensive tests. This scheduler would also feed multiple Builders.

Many schedulers can be configured to wait a while after seeing a source-code change - this is the tree stable timer. The timer allows multiple commits to be "batched" together. This is particularly useful in distributed version control systems, where a developer may push a long sequence of changes all at once. To save resources, it's often desirable only to test the most recent change.

Schedulers can also filter out the changes they are interested in, based on a number of criteria. For example, a scheduler that only builds documentation might skip any changes that do not affect the documentation. Schedulers can also filter on the branch to which a commit was made.

There is some support for configuring dependencies between builds - for example, you may want to build packages only for revisions which pass all of the unit tests. This support is under active development in Buildbot, and is referred to as "build coordination".

Periodic builds (those which are run every N seconds rather than after new Changes arrive) are triggered by a special Periodic scheduler.

Each scheduler creates and submits BuildSet objects to the BuildMaster, which is then responsible for making sure the individual BuildRequests are delivered to the target Builders.

Scheduler instances are activated by placing them in the schedulers list in the buildmaster config file. Each scheduler must have a unique name.

2.3.5. BuildSets

A BuildSet is the name given to a set of Builds that all compile/test the same version of the tree on multiple Builders. In general, all these component Builds will perform the same sequence of Steps, using the same source code, but on different platforms or against a different set of libraries.

The BuildSet is tracked as a single unit, which fails if any of the component Builds have failed, and therefore can succeed only if all of the component Builds have succeeded. There are two kinds of status notification messages that can be emitted for a BuildSet: the firstFailure type (which fires as soon as we know the BuildSet will fail), and the Finished type (which fires once the BuildSet has completely finished, regardless of whether the overall set passed or failed).

A BuildSet is created with set of one or more source stamp tuples of (branch, revision, changes, patch), some of which may be None, and a list of Builders on which it is to be run. They are then given to the BuildMaster, which is responsible for creating a separate BuildRequest for each Builder.

There are a couple of different likely values for the SourceStamp:

(revision=None, changes=CHANGES, patch=None)
This is a SourceStamp used when a series of Changes have triggered a build. The VC step will attempt to check out a tree that contains CHANGES (and any changes that occurred before CHANGES, but not any that occurred after them.)
(revision=None, changes=None, patch=None)
This builds the most recent code on the default branch. This is the sort of SourceStamp that would be used on a Build that was triggered by a user request, or a Periodic scheduler. It is also possible to configure the VC Source Step to always check out the latest sources rather than paying attention to the Changes in the SourceStamp, which will result in same behavior as this.
(branch=BRANCH, revision=None, changes=None, patch=None)
This builds the most recent code on the given BRANCH. Again, this is generally triggered by a user request or a Periodic scheduler.
(revision=REV, changes=None, patch=(LEVEL, DIFF, SUBDIR_ROOT))
This checks out the tree at the given revision REV, then applies a patch (using patch -pLEVEL <DIFF) from inside the relative directory SUBDIR_ROOT. Item SUBDIR_ROOT is optional and defaults to the builder working directory. The try command creates this kind of SourceStamp. If patch is None, the patching step is bypassed.

The buildmaster is responsible for turning the BuildSet into a set of BuildRequest objects and queueing them on the appropriate Builders.

2.3.6. BuildRequests

A BuildRequest is a request to build a specific set of source code (specified by one ore more source stamps) on a single Builder. Each Builder runs the BuildRequest as soon as it can (i.e. when an associated worker becomes free). BuildRequests are prioritized from oldest to newest, so when a worker becomes free, the Builder with the oldest BuildRequest is run.

The BuildRequest contains one SourceStamp specification per codebase. The actual process of running the build (the series of Steps that will be executed) is implemented by the Build object. In the future this might be changed, to have the Build define what gets built, and a separate BuildProcess (provided by the Builder) to define how it gets built.

The BuildRequest may be mergeable with other compatible BuildRequests. Builds that are triggered by incoming Changes will generally be mergeable. Builds that are triggered by user requests are generally not, unless they are multiple requests to build the latest sources of the same branch. A merge of buildrequests is performed per codebase, thus on changes having the same codebase.

2.3.7. Builders

The Buildmaster runs a collection of Builders, each of which handles a single type of build (e.g. full versus quick), on one or more workers. Builders serve as a kind of queue for a particular type of build. Each Builder gets a separate column in the waterfall display. In general, each Builder runs independently (although various kinds of interlocks can cause one Builder to have an effect on another).

Each builder is a long-lived object which controls a sequence of Builds. Each Builder is created when the config file is first parsed, and lives forever (or rather until it is removed from the config file). It mediates the connections to the workers that do all the work, and is responsible for creating the Build objects - Builds.

Each builder gets a unique name, and the path name of a directory where it gets to do all its work (there is a buildmaster-side directory for keeping status information, as well as a worker-side directory where the actual checkout/compile/test commands are executed).

2.3.8. Build Factories

A builder also has a BuildFactory, which is responsible for creating new Build instances: because the Build instance is what actually performs each build, choosing the BuildFactory is the way to specify what happens each time a build is done (Builds).

2.3.9. Workers

Each builder is associated with one of more Workers. A builder which is used to perform Mac OS X builds (as opposed to Linux or Solaris builds) should naturally be associated with a Mac worker.

If multiple workers are available for any given builder, you will have some measure of redundancy: in case one worker goes offline, the others can still keep the Builder working. In addition, multiple workers will allow multiple simultaneous builds for the same Builder, which might be useful if you have a lot of forced or try builds taking place.

If you use this feature, it is important to make sure that the workers are all, in fact, capable of running the given build. The worker hosts should be configured similarly, otherwise you will spend a lot of time trying (unsuccessfully) to reproduce a failure that only occurs on some of the workers and not the others. Different platforms, operating systems, versions of major programs or libraries, all these things mean you should use separate Builders.

2.3.10. Builds

A build is a single compile or test run of a particular version of the source code, and is comprised of a series of steps. It is ultimately up to you what constitutes a build, but for compiled software it is generally the checkout, configure, make, and make check sequence. For interpreted projects like Python modules, a build is generally a checkout followed by an invocation of the bundled test suite.

A BuildFactory describes the steps a build will perform. The builder which starts a build uses its configured build factory to determine the build's steps.

2.3.11. Users

Buildbot has a somewhat limited awareness of users. It assumes the world consists of a set of developers, each of whom can be described by a couple of simple attributes. These developers make changes to the source code, causing builds which may succeed or fail.

Users also may have different levels of authorization when issuing Buildbot commands, such as forcing a build from the web interface or from an IRC channel.

Each developer is primarily known through the source control system. Each Change object that arrives is tagged with a who field that typically gives the account name (on the repository machine) of the user responsible for that change. This string is displayed on the HTML status pages and in each Build's blamelist.

To do more with the User than just refer to them, this username needs to be mapped into an address of some sort. The responsibility for this mapping is left up to the status module which needs the address. In the future, the responsibility for managing users will be transferred to User Objects.

The who fields in git Changes are used to create User Objects, which allows for more control and flexibility in how Buildbot manages users. User Objects

User Objects allow Buildbot to better manage users throughout its various interactions with users (see Change Sources and Reporters). The User Objects are stored in the Buildbot database and correlate the various attributes that a user might have: irc, Git, etc.


Incoming Changes all have a who attribute attached to them that specifies which developer is responsible for that Change. When a Change is first rendered, the who attribute is parsed and added to the database if it doesn't exist or checked against an existing user. The who attribute is formatted in different ways depending on the version control system that the Change came from.

who attributes take the form Full Name <Email>.
who attributes are of the form Username.
who attributes are free-form strings, but usually adhere to similar conventions as git attributes (Full Name <Email>).
who attributes are of the form Username.
who attributes contain an Email and may also include a Full Name like git attributes.
who attributes are free-form strings like hg, and can include a Username, Email, and/or Full Name.

For managing users manually, use the buildbot user command, which allows you to add, remove, update, and show various attributes of users in the Buildbot database (see Command-line Tool).


Correlating the various bits and pieces that Buildbot views as users also means that one attribute of a user can be translated into another. This provides a more complete view of users throughout Buildbot.

One such use is being able to find email addresses based on a set of Builds to notify users through the MailNotifier. This process is explained more clearly in Email Addresses.

Another way to utilize User Objects is through UsersAuth for web authentication. To use UsersAuth, you need to set a bb_username and bb_password via the buildbot user command line tool to check against. The password will be encrypted before storing in the database along with other user attributes. Doing Things With Users

Each change has a single user who is responsible for it. Most builds have a set of changes: the build generally represents the first time these changes have been built and tested by the Buildbot. The build has a blamelist that is the union of the users responsible for all the build's changes. If the build was created by a Try Schedulers this list will include the submitter of the try job, if known.

The build provides a list of users who are interested in the build -- the interested users. Usually this is equal to the blamelist, but may also be expanded, e.g., to include the current build sherrif or a module's maintainer.

If desired, the buildbot can notify the interested users until the problem is resolved. Email Addresses

The MailNotifier is a status target which can send email about the results of each build. It accepts a static list of email addresses to which each message should be delivered, but it can also be configured to send mail to the Build's Interested Users. To do this, it needs a way to convert User names into email addresses.

For many VC systems, the User Name is actually an account name on the system which hosts the repository. As such, turning the name into an email address is a simple matter of appending Some projects use other kinds of mappings (for example the preferred email address may be at despite the repository host being named, and some VC systems have full separation between the concept of a user and that of an account on the repository host (like Perforce). Some systems (like Git) put a full contact email address in every change.

To convert these names to addresses, the MailNotifier uses an EmailLookup object. This provides a getAddress method which accepts a name and (eventually) returns an address. The default MailNotifier module provides an EmailLookup which simply appends a static string, configurable when the notifier is created. To create more complex behaviors (perhaps using an LDAP lookup, or using finger on a central host to determine a preferred address for the developer), provide a different object as the lookup argument.

If an EmailLookup object isn't given to the MailNotifier, the MailNotifier will try to find emails through User Objects. This will work the same as if an EmailLookup object was used if every user in the Build's Interested Users list has an email in the database for them. If a user whose change led to a Build doesn't have an email attribute, that user will not receive an email. If extraRecipients is given, those users are still sent mail when the EmailLookup object is not specified.

In the future, when the Problem mechanism has been set up, the Buildbot will need to send mail to arbitrary Users. It will do this by locating a MailNotifier-like object among all the buildmaster's status targets, and asking it to send messages to various Users. This means the User-to-address mapping only has to be set up once, in your MailNotifier, and every email message the buildbot emits will take advantage of it. IRC Nicknames

Like MailNotifier, the buildbot.status.words.IRC class provides a status target which can announce the results of each build. It also provides an interactive interface by responding to online queries posted in the channel or sent as private messages.

In the future, the buildbot can be configured map User names to IRC nicknames, to watch for the recent presence of these nicknames, and to deliver build status messages to the interested parties. Like MailNotifier does for email addresses, the IRC object will have an IRCLookup which is responsible for nicknames. The mapping can be set up statically, or it can be updated by online users themselves (by claiming a username with some kind of buildbot: i am user warner commands).

Once the mapping is established, the rest of the buildbot can ask the IRC object to send messages to various users. It can report on the likelihood that the user saw the given message (based upon how long the user has been inactive on the channel), which might prompt the Problem Hassler logic to send them an email message instead.

These operations and authentication of commands issued by particular nicknames will be implemented in User Objects.

2.3.12. Build Properties

Each build has a set of Build Properties, which can be used by its build steps to modify their actions. These properties, in the form of key-value pairs, provide a general framework for dynamically altering the behavior of a build based on its circumstances.

Properties form a simple kind of variable in a build. Some properties are set when the build starts, and properties can be changed as a build progresses -- properties set or changed in one step may be accessed in subsequent steps. Property values can be numbers, strings, lists, or dictionaries - basically, anything that can be represented in JSON.

Properties are very flexible, and can be used to implement all manner of functionality. Here are some examples:

Most Source steps record the revision that they checked out in the got_revision property. A later step could use this property to specify the name of a fully-built tarball, dropped in an easily-accessible directory for later testing.


In builds with more than one codebase, the got_revision property is a dictionary, keyed by codebase.

Some projects want to perform nightly builds as well as building in response to committed changes. Such a project would run two schedulers, both pointing to the same set of builders, but could provide an is_nightly property so that steps can distinguish the nightly builds, perhaps to run more resource-intensive tests.

Some projects have different build processes on different systems. Rather than create a build factory for each worker, the steps can use worker properties to identify the unique aspects of each worker and adapt the build process dynamically.

2.3.13. Multiple-Codebase Builds

What if an end-product is composed of code from several codebases? Changes may arrive from different repositories within the tree-stable-timer period. Buildbot will not only use the source-trees that contain changes but also needs the remaining source-trees to build the complete product.

For this reason a Scheduler can be configured to base a build on a set of several source-trees that can (partly) be overridden by the information from incoming Changes.

As described above, the source for each codebase is identified by a source stamp, containing its repository, branch and revision. A full build set will specify a source stamp set describing the source to use for each codebase.

Configuring all of this takes a coordinated approach. A complete multiple repository configuration consists of:

a codebase generator

Every relevant change arriving from a VC must contain a codebase. This is done by a codebaseGenerator that is defined in the configuration. Most generators examine the repository of a change to determine its codebase, using project-specific rules.

some schedulers

Each scheduler has to be configured with a set of all required codebases to build a product. These codebases indicate the set of required source-trees. In order for the scheduler to be able to produce a complete set for each build, the configuration can give a default repository, branch, and revision for each codebase. When a scheduler must generate a source stamp for a codebase that has received no changes, it applies these default values.

multiple source steps - one for each codebase

A Builders's build factory must include a source step for each codebase. Each of the source steps has a codebase attribute which is used to select an appropriate source stamp from the source stamp set for a build. This information comes from the arrived changes or from the scheduler's configured default values.


Each source step has to have its own workdir set in order for the checkout to be done for each codebase in its own directory.


Ensure you specify the codebase within your source step's Interpolate() calls (ex. http://.../svn/%(src:codebase:branch)s). See Interpolate for details.


Defining a codebaseGenerator that returns non-empty (not '') codebases will change the behavior of all the schedulers.

2.3.14. Multimaster

blockdiag LoadBalancer LoadBalancerUI Worker1 WorkerN User1 User2 Master1 Master2 MasterUI1 MasterUI2 database

Buildbot supports interconnection of several masters. This has to be done through a multi-master enabled message queue backend. As of now the only one supported is wamp and see wamp

There are then several strategy for introducing multimaster in your buildbot infra. A simple way to say it is by adding the concept of symmetrics and asymmetrics multimaster (like there is SMP and AMP for multi core CPUs)

Symmetric multimaster is when each master share the exact same configuration. They run the same builders, same schedulers, same everything, the only difference is that workers are connected evenly between the masters (by any means (e.g. DNS load balancing, etc)) Symmetric multimaster is good to use to scale buildbot horizontally.

Asymmetric multimaster is when each master have different configuration. Each master may have a specific responsibility (e.g schedulers, set of builder, UI). This was more how you did in 0.8, also because of its own technical limitations. A nice feature of asymmetric multimaster is that you can have the UI only handled by some masters.

Separating the UI from the controlling will greatly help in the performance of the UI, because badly written BuildSteps?? can stall the reactor for several seconds.

The fanciest configuration would probably be a symmetric configuration for everything but the UI. You would scale the number of UI master according to your number of UI users, and scale the number of engine masters to the number of workers.

Depending on your workload and size of master host, it is probably a good idea to start thinking of multimaster starting from a hundred workers connected.

Multimaster can also be used for high availability, and seamless upgrade of configuration code. Complex configuration indeed requires sometimes to restart the master to reload custom steps or code, or just to upgrade the upstream buildbot version.

In this case, you will implement following procedure:

  • Start new master(s) with new code and configuration.
  • Send a graceful shutdown to the old master(s).
  • New master(s) will start taking the new jobs, while old master(s) will just finish managing the running builds.
  • As an old master is finishing the running builds, it will drop the connections from the workers, who will then reconnect automatically, and by the mean of load balancer will get connected to a new master to run new jobs.

As buildbot nine has been designed to allow such procedure, it has not been implemented in production yet as we know. There is probably a new REST api needed in order to graceful shutdown a master, and the details of gracefully dropping the connection to the workers to be sorted out.

2.4. Configuration

The following sections describe the configuration of the various Buildbot components. The information available here is sufficient to create basic build and test configurations, and does not assume great familiarity with Python.

In more advanced Buildbot configurations, Buildbot acts as a framework for a continuous-integration application. The next section, Customization, describes this approach, with frequent references into the development documentation.

2.4.1. Configuring Buildbot

The buildbot's behavior is defined by the config file, which normally lives in the master.cfg file in the buildmaster's base directory (but this can be changed with an option to the buildbot create-master command). This file completely specifies which Builders are to be run, which workers they should use, how Changes should be tracked, and where the status information is to be sent. The buildmaster's buildbot.tac file names the base directory; everything else comes from the config file.

A sample config file was installed for you when you created the buildmaster, but you will need to edit it before your buildbot will do anything useful.

This chapter gives an overview of the format of this file and the various sections in it. You will need to read the later chapters to understand how to fill in each section properly. Config File Format

The config file is, fundamentally, just a piece of Python code which defines a dictionary named BuildmasterConfig, with a number of keys that are treated specially. You don't need to know Python to do basic configuration, though, you can just copy the syntax of the sample file. If you are comfortable writing Python code, however, you can use all the power of a full programming language to achieve more complicated configurations.

The BuildmasterConfig name is the only one which matters: all other names defined during the execution of the file are discarded. When parsing the config file, the Buildmaster generally compares the old configuration with the new one and performs the minimum set of actions necessary to bring the buildbot up to date: Builders which are not changed are left untouched, and Builders which are modified get to keep their old event history.

The beginning of the master.cfg file typically starts with something like:

BuildmasterConfig = c = {}

Therefore a config key like change_source will usually appear in master.cfg as c['change_source'].

See Buildmaster Configuration Index for a full list of BuildMasterConfig keys.

Basic Python Syntax

The master configuration file is interpreted as Python, allowing the full flexibility of the language. For the configurations described in this section, a detailed knowledge of Python is not required, but the basic syntax is easily described.

Python comments start with a hash character #, tuples are defined with (parenthesis, pairs), and lists (arrays) are defined with [square, brackets]. Tuples and lists are mostly interchangeable. Dictionaries (data structures which map keys to values) are defined with curly braces: {'key1': value1, 'key2': value2}. Function calls (and object instantiation) can use named parameters, like steps.ShellCommand(command=["trial", "pyflakes"]).

The config file starts with a series of import statements, which make various kinds of Steps and Status targets available for later use. The main BuildmasterConfig dictionary is created, then it is populated with a variety of keys, described section-by-section in subsequent chapters. Predefined Config File Symbols

The following symbols are automatically available for use in the configuration file.


the base directory for the buildmaster. This string has not been expanded, so it may start with a tilde. It needs to be expanded before use. The config file is located in:

os.path.expanduser(os.path.join(basedir, 'master.cfg'))
the absolute path of the config file. The config file's directory is located in os.path.dirname(__file__). Testing the Config File

To verify that the config file is well-formed and contains no deprecated or invalid elements, use the checkconfig command, passing it either a master directory or a config file.

% buildbot checkconfig master.cfg
Config file is good!
# or
% buildbot checkconfig /tmp/masterdir
Config file is good!

If the config file has deprecated features (perhaps because you've upgraded the buildmaster and need to update the config file to match), they will be announced by checkconfig. In this case, the config file will work, but you should really remove the deprecated items and use the recommended replacements instead:

% buildbot checkconfig master.cfg
/usr/lib/python2.4/site-packages/buildbot/ DeprecationWarning: c['sources'] is
deprecated as of 0.7.6 and will be removed by 0.8.0 . Please use c['change_source'] instead.
Config file is good!

If you have errors in your configuration file, checkconfig will let you know:

% buildbot checkconfig master.cfg
Configuration Errors:
c['workers'] must be a list of Worker instances
no workers are configured
builder 'smoketest' uses unknown workers 'linux-002'

If the config file is simply broken, that will be caught too:

% buildbot checkconfig master.cfg
error while parsing config file:
Traceback (most recent call last):
File "/home/buildbot/master/bin/buildbot", line 4, in <module>
File "/home/buildbot/master/buildbot/scripts/", line 1358, in run
    if not doCheckConfig(so):
File "/home/buildbot/master/buildbot/scripts/", line 1079, in doCheckConfig
    return cl.load(quiet=quiet)
File "/home/buildbot/master/buildbot/scripts/", line 29, in load
    self.basedir, self.configFileName)
--- <exception caught here> ---
File "/home/buildbot/master/buildbot/", line 147, in loadConfig
    exec f in localDict
exceptions.SyntaxError: invalid syntax (master.cfg, line 52)
Configuration Errors:
error while parsing config file: invalid syntax (master.cfg, line 52) (traceback in logfile) Loading the Config File

The config file is only read at specific points in time. It is first read when the buildmaster is launched.


If the configuration is invalid, the master will display the errors in the console output, but will not exit.

Reloading the Config File (reconfig)

If you are on the system hosting the buildmaster, you can send a SIGHUP signal to it: the buildbot tool has a shortcut for this:

buildbot reconfig BASEDIR

This command will show you all of the lines from twistd.log that relate to the reconfiguration. If there are any problems during the config-file reload, they will be displayed in these lines.

When reloading the config file, the buildmaster will endeavor to change as little as possible about the running system. For example, although old status targets may be shut down and new ones started up, any status targets that were not changed since the last time the config file was read will be left running and untouched. Likewise any Builders which have not been changed will be left running. If a Builder is modified (say, the build process is changed) while a Build is currently running, that Build will keep running with the old process until it completes. Any previously queued Builds (or Builds which get queued after the reconfig) will use the new process.


Buildbot's reconfiguration system is fragile for a few difficult-to-fix reasons:

  • Any modules imported by the configuration file are not automatically reloaded. Python modules such as may help here, but reloading modules is fraught with subtleties and difficult-to-decipher failure cases.
  • During the reconfiguration, active internal objects are divorced from the service hierarchy, leading to tracebacks in the web interface and other components. These are ordinarily transient, but with HTTP connection caching (either by the browser or an intervening proxy) they can last for a long time.
  • If the new configuration file is invalid, it is possible for Buildbot's internal state to be corrupted, leading to undefined results. When this occurs, it is best to restart the master.
  • For more advanced configurations, it is impossible for Buildbot to tell if the configuration for a Builder or Scheduler has changed, and thus the Builder or Scheduler will always be reloaded. This occurs most commonly when a callable is passed as a configuration parameter.

The bbproto project (at may help to construct large (multi-file) configurations which can be effectively reloaded and reconfigured.

2.4.2. Global Configuration

The keys in this section affect the operations of the buildmaster globally. Database Specification

Buildbot requires a connection to a database to maintain certain state information, such as tracking pending build requests. In the default configuration Buildbot uses a file-based SQLite database, stored in the state.sqlite file of the master's base directory. Override this configuration with the db_url parameter.

Buildbot accepts a database configuration in a dictionary named db. All keys are optional:

c['db'] = {
    'db_url' : 'sqlite:///state.sqlite',

The db_url key indicates the database engine to use. The format of this parameter is completely documented at, but is generally of the form:


These parameters can be specified directly in the configuration dictionary, as c['db_url'] and c['db_poll_interval'], although this method is deprecated.

The following sections give additional information for particular database backends:


For sqlite databases, since there is no host and port, relative paths are specified with sqlite:/// and absolute paths with sqlite:////. Examples:

c['db_url'] = "sqlite:///state.sqlite"

SQLite requires no special configuration.

c['db_url'] = "mysql://"

The max_idle argument for MySQL connections is unique to Buildbot, and should be set to something less than the wait_timeout configured for your server. This controls the SQLAlchemy pool_recycle parameter, which defaults to no timeout. Setting this parameter ensures that connections are closed and re-opened after the configured amount of idle time. If you see errors such as _mysql_exceptions.OperationalError: (2006, 'MySQL server has gone away'), this means your max_idle setting is probably too high. show global variables like 'wait_timeout'; will show what the currently configured wait_timeout is on your MySQL server.

When using MySQL 5.x, if you see errors such as BLOB, TEXT, GEOMETRY or JSON column state_string can not have a default value make sure to add sql_mode='MYSQL40' in your configuration cnf file.

Buildbot requires use_unique=True and charset=utf8, and will add them automatically, so they do not need to be specified in db_url.

MySQL defaults to the MyISAM storage engine, but this can be overridden with the storage_engine URL argument.

c['db_url'] = "postgresql://username@hostname/dbname"

PosgreSQL requires no special configuration. MQ Specification

Buildbot uses a message-queueing system to handle communication within the master. Messages are used to indicate events within the master, and components that are interested in those events arrange to receive them.

The message queueing implementation is configured as a dictionary in the mq option. The type key describes the type of MQ implemetation to be used. Note that the implementation type cannot be changed in a reconfig.

The available implemenetation types are described in the following sections.

c['mq'] = {
    'type' : 'simple',
    'debug' : False,

This is the default MQ implementation. Similar to SQLite, it has no additional software dependencies, but does not support multi-master mode.

Note that this implementation also does not support message persistence across a restart of the master. For example, if a change is received, but the master shuts down before the schedulers can create build requests for it, then those schedulers will not be notified of the change when the master starts again.

The debug key, which defaults to False, can be used to enable logging of every message produced on this master.

c['mq'] = {
    'type' : 'wamp',
    'router_url': 'ws://url/to/crossbar'
    'realm': 'buildbot'
    'debug' : False,
    'debug_websockets' : False,
    'debug_lowlevel' : False,

This is a MQ implementation using wamp protocol. This implementation uses Python Autobahn wamp client library, and is fully asynchronous (no use of threads) To use this implementation, you need a wamp router like Crossbar. The implementation does not yet support wamp authentication yet. This MQ allows buildbot to run in multi-master mode.

Note that this implementation also does not support message persistence across a restart of the master. For example, if a change is received, but the master shuts down before the schedulers can create build requests for it, then those schedulers will not be notified of the change when the master starts again.

router_url key is mandatory, and should point to your router websocket url. Buildbot is only supporting wamp over websocket, which is a sub-protocol of http. SSL is supported using wss:// instead of ws://. You must use a router with very reliable connection to the master. If for some reason, the wamp connection is lost, then the master will stop, and should be restarted via a process manager.

realm key is optional and defaults to buildbot, and configures the wamp realm to use for your buildbot messages.

The debug key, which defaults to False, can be used to enable logging of every message produced on this master. debug_websocket and debug_lowlevel, enable more debug logs in autobahn. Multi-master mode

Normally buildbot operates using a single master process that uses the configured database to save state.

It is possible to configure buildbot to have multiple master processes that share state in the same database. This has been well tested using a MySQL database. There are several benefits of Multi-master mode:

  • You can have large numbers of workers handling the same queue of build requests. A single master can only handle so many workers (the number is based on a number of factors including type of builds, number of builds, and master and worker IO and CPU capacity--there is no fixed formula). By adding another master which shares the queue of build requests, you can attach more workers to this additional master, and increase your build throughput.
  • You can shut one master down to do maintenance, and other masters will continue to do builds.

State that is shared in the database includes:

  • List of changes
  • Scheduler names and internal state
  • Build requests, including the builder name

Because of this shared state, you are strongly encouraged to:

  • Ensure that each named scheduler runs on only one master. If the same scheduler runs on multiple masters, it will trigger duplicate builds and may produce other undesirable behaviors.
  • Ensure builder names are unique for a given build factory implementation. You can have the same builder name configured on many masters, but if the build factories differ, you will get different results depending on which master claims the build.

One suggested configuration is to have one buildbot master configured with just the scheduler and change sources; and then other masters configured with just the builders.

To enable multi-master mode in this configuration, you will need to set the multiMaster option so that buildbot doesn't warn about missing schedulers or builders.

# Enable multiMaster mode; disables warnings about unknown builders and
# schedulers
c['multiMaster'] = True
# Check for new build requests every 60 seconds
c['db'] = {
    'db_url' : 'mysql://...',
} Site Definitions

Three basic settings describe the buildmaster in status reports:

c['title'] = "Buildbot"
c['titleURL'] = ""

title is a short string that will appear at the top of this buildbot installation's home page (linked to the titleURL).

titleURL is a URL string that must end with a slash (/). HTML status displays will show title as a link to titleURL. This URL is often used to provide a link from buildbot HTML pages to your project's home page.

The buildbotURL string should point to the location where the buildbot's internal web server is visible. This URL must end with a slash (/).

When status notices are sent to users (e.g., by email or over IRC), buildbotURL will be used to create a URL to the specific build or problem that they are being notified about. Log Handling
c['logCompressionLimit'] = 16384
c['logCompressionMethod'] = 'gz'
c['logMaxSize'] = 1024*1024 # 1M
c['logMaxTailSize'] = 32768
c['logEncoding'] = 'utf-8'

The logCompressionLimit enables compression of build logs on disk for logs that are bigger than the given size, or disables that completely if set to False. The default value is 4096, which should be a reasonable default on most file systems. This setting has no impact on status plugins, and merely affects the required disk space on the master for build logs.

The logCompressionMethod controls what type of compression is used for build logs. The default is 'gz', and the other valid option are 'raw' (no compression), 'gz' or 'lz4' (required lz4 package).

Please find below some stats extracted from 50x "Pyflakes" runs (results may differ according to log type).

Space saving details
compression raw log size compressed log size space saving compression speed
bz2 2.981 MB 0.603 MB 79.77% 3.433 MB/s
gz 2.981 MB 0.568 MB 80.95% 6.604 MB/s
lz4 2.981 MB 0.844 MB 71.68% 77.668 MB/s

The logMaxSize parameter sets an upper limit (in bytes) to how large logs from an individual build step can be. The default value is None, meaning no upper limit to the log size. Any output exceeding logMaxSize will be truncated, and a message to this effect will be added to the log's HEADER channel.

If logMaxSize is set, and the output from a step exceeds the maximum, the logMaxTailSize parameter controls how much of the end of the build log will be kept. The effect of setting this parameter is that the log will contain the first logMaxSize bytes and the last logMaxTailSize bytes of output. Don't set this value too high, as the the tail of the log is kept in memory.

The logEncoding parameter specifies the character encoding to use to decode bytestrings provided as logs. It defaults to utf-8, which should work in most cases, but can be overridden if necessary. In extreme cases, a callable can be specified for this parameter. It will be called with byte strings, and should return the corresponding Unicode string.

This setting can be overridden for a single build step with the logEncoding step parameter. It can also be overridden for a single log file by passing the logEncoding parameter to addLog. Data Lifetime
c['changeHorizon'] = 200
c['buildHorizon'] = 100
c['eventHorizon'] = 50
c['logHorizon'] = 40
c['buildCacheSize'] = 15

Buildbot stores historical information on disk in the form of "Pickle" files and compressed logfiles. In a large installation, these can quickly consume disk space, yet in many cases developers never consult this historical information.

The changeHorizon key determines how many changes the master will keep a record of. One place these changes are displayed is on the waterfall page. This parameter defaults to 0, which means keep all changes indefinitely.

The buildHorizon specifies the minimum number of builds for each builder which should be kept on disk. The eventHorizon specifies the minimum number of events to keep--events mostly describe connections and disconnections of workers, and are seldom helpful to developers. The logHorizon gives the minimum number of builds for which logs should be maintained; this parameter must be less than or equal to buildHorizon. Builds older than logHorizon but not older than buildHorizon will maintain their overall status and the status of each step, but the logfiles will be deleted.

c['caches'] = {
    'Changes' : 100,     # formerly c['changeCacheSize']
    'Builds' : 500,      # formerly c['buildCacheSize']
    'chdicts' : 100,
    'BuildRequests' : 10,
    'SourceStamps' : 20,
    'ssdicts' : 20,
    'objectids' : 10,
    'usdicts' : 100,

The caches configuration key contains the configuration for Buildbot's in-memory caches. These caches keep frequently-used objects in memory to avoid unnecessary trips to the database or to pickle files. Caches are divided by object type, and each has a configurable maximum size.

The default size for each cache is 1, except where noted below. A value of 1 allows Buildbot to make a number of optimizations without consuming much memory. Larger, busier installations will likely want to increase these values.

The available caches are:


the number of change objects to cache in memory. This should be larger than the number of changes that typically arrive in the span of a few minutes, otherwise your schedulers will be reloading changes from the database every time they run. For distributed version control systems, like Git or Hg, several thousand changes may arrive at once, so setting this parameter to something like 10000 isn't unreasonable.

This parameter is the same as the deprecated global parameter changeCacheSize. Its default value is 10.


The buildCacheSize parameter gives the number of builds for each builder which are cached in memory. This number should be larger than the number of builds required for commonly-used status displays (the waterfall or grid views), so that those displays do not miss the cache on a refresh.

This parameter is the same as the deprecated global parameter buildCacheSize. Its default value is 15.

The number of rows from the changes table to cache in memory. This value should be similar to the value for Changes.
The number of BuildRequest objects kept in memory. This number should be higher than the typical number of outstanding build requests. If the master ordinarily finds jobs for BuildRequests immediately, you may set a lower value.
the number of SourceStamp objects kept in memory. This number should generally be similar to the number BuildRequesets.
The number of rows from the sourcestamps table to cache in memory. This value should be similar to the value for SourceStamps.
The number of object IDs - a means to correlate an object in the Buildbot configuration with an identity in the database--to cache. In this version, object IDs are not looked up often during runtime, so a relatively low value such as 10 is fine.

The number of rows from the users table to cache in memory. Note that for a given user there will be a row for each attribute that user has.

c['buildCacheSize'] = 15 Merging Build Requests
c['collapseRequests'] = True

This is a global default value for builders' collapseRequests parameter, and controls the merging of build requests.

This parameter can be overridden on a per-builder basis. See Collapsing Build Requests for the allowed values for this parameter. Prioritizing Builders
def prioritizeBuilders(buildmaster, builders):
c['prioritizeBuilders'] = prioritizeBuilders

By default, buildbot will attempt to start builds on builders in order, beginning with the builder with the oldest pending request. Customize this behavior with the prioritizeBuilders configuration key, which takes a callable. See Builder Priority Functions for details on this callable.

This parameter controls the order that the build master can start builds, and is useful in situations where there is resource contention between builders, e.g., for a test database. It does not affect the order in which a builder processes the build requests in its queue. For that purpose, see Prioritizing Builds. Setting the PB Port for Workers
c['protocols'] = {"pb": {"port": 10000}}

The buildmaster will listen on a TCP port of your choosing for connections from workers. It can also use this port for connections from remote Change Sources, status clients, and debug tools. This port should be visible to the outside world, and you'll need to tell your worker admins about your choice.

It does not matter which port you pick, as long it is externally visible; however, you should probably use something larger than 1024, since most operating systems don't allow non-root processes to bind to low-numbered ports. If your buildmaster is behind a firewall or a NAT box of some sort, you may have to configure your firewall to permit inbound connections to this port.

c['protocols']['pb']['port'] is a strports specification string, defined in the twisted.application.strports module (try pydoc twisted.application.strports to get documentation on the format).

This means that you can have the buildmaster listen on a localhost-only port by doing:

c['protocols'] = {"pb": {"port": "tcp:10000:interface="}}

This might be useful if you only run workers on the same machine, and they are all configured to contact the buildmaster at localhost:10000.


In Buildbot versions <=0.8.8 you might see slavePortnum option. This option contains same value as c['protocols']['pb']['port'] but not recomended to use. Defining Global Properties

The properties configuration key defines a dictionary of properties that will be available to all builds started by the buildmaster:

c['properties'] = {
    'Widget-version' : '1.2',
    'release-stage' : 'alpha'
} Manhole

If you set manhole to an instance of one of the classes in buildbot.manhole, you can telnet or ssh into the buildmaster and get an interactive Python shell, which may be useful for debugging buildbot internals. It is probably only useful for buildbot developers. It exposes full access to the buildmaster's account (including the ability to modify and delete files), so it should not be enabled with a weak or easily guessable password.

There are three separate Manhole classes. Two of them use SSH, one uses unencrypted telnet. Two of them use a username+password combination to grant access, one of them uses an SSH-style authorized_keys file which contains a list of ssh public keys.


Using any Manhole requires that pycrypto and pyasn1 be installed. These are not part of the normal Buildbot dependencies.

You construct this with the name of a file that contains one SSH public key per line, just like ~/.ssh/authorized_keys. If you provide a non-absolute filename, it will be interpreted relative to the buildmaster's base directory.
This one accepts SSH connections but asks for a username and password when authenticating. It accepts only one such pair.
This accepts regular unencrypted telnet connections, and asks for a username/password pair before providing access. Because this username/password is transmitted in the clear, and because Manhole access to the buildmaster is equivalent to granting full shell privileges to both the buildmaster and all the workers (and to all accounts which then run code produced by the workers), it is highly recommended that you use one of the SSH manholes instead.
# some examples:
from buildbot.plugins import util
c['manhole'] = util.AuthorizedKeysManhole(1234, "authorized_keys")
c['manhole'] = util.PasswordManhole(1234, "alice", "mysecretpassword")
c['manhole'] = util.TelnetManhole(1234, "bob", "snoop_my_password_please")

The Manhole instance can be configured to listen on a specific port. You may wish to have this listening port bind to the loopback interface (sometimes known as lo0, localhost, or to restrict access to clients which are running on the same host.

from buildbot.plugins import util
c['manhole'] = util.PasswordManhole("tcp:9999:interface=","admin","passwd")

To have the Manhole listen on all interfaces, use "tcp:9999" or simply 9999. This port specification uses twisted.application.strports, so you can make it listen on SSL or even UNIX-domain sockets if you want.

Note that using any Manhole requires that the TwistedConch package be installed.

The buildmaster's SSH server will use a different host key than the normal sshd running on a typical unix host. This will cause the ssh client to complain about a host key mismatch, because it does not realize there are two separate servers running on the same host. To avoid this, use a clause like the following in your .ssh/config file:

Host remotehost-buildbot
HostName remotehost
HostKeyAlias remotehost-buildbot
Port 9999
# use 'user' if you use PasswordManhole and your name is not 'admin'.
# if you use AuthorizedKeysManhole, this probably doesn't matter.
User admin
Using Manhole

After you have connected to a manhole instance, you will find yourself at a Python prompt. You have access to two objects: master (the BuildMaster) and status (the master's Status object). Most interesting objects on the master can be reached from these two objects.

To aid in navigation, the show method is defined. It displays the non-method attributes of an object.

A manhole session might look like:

>>> show(master)
data attributes of <buildbot.master.BuildMaster instance at 0x7f7a4ab7df38>
                       basedir : '/home/dustin/code/buildbot/t/buildbot/'...
                     botmaster : <type 'instance'>
                buildCacheSize : None
                  buildHorizon : None
                   buildbotURL : http://localhost:8010/
               changeCacheSize : None
                    change_svc : <type 'instance'>
                configFileName : master.cfg
                            db : <class 'buildbot.db.connector.DBConnector'>
                        db_url : sqlite:///state.sqlite
>>> show(['win32'])
data attributes of <Builder ''builder'' at 48963528>
>>> win32 = _
>>> win32.category = 'w32' Metrics Options
c['metrics'] = dict(log_interval=10, periodic_interval=10)

metrics can be a dictionary that configures various aspects of the metrics subsystem. If metrics is None, then metrics collection, logging and reporting will be disabled.

log_interval determines how often metrics should be logged to twistd.log. It defaults to 60s. If set to 0 or None, then logging of metrics will be disabled. This value can be changed via a reconfig.

periodic_interval determines how often various non-event based metrics are collected, such as memory usage, uncollectable garbage, reactor delay. This defaults to 10s. If set to 0 or None, then periodic collection of this data is disabled. This value can also be changed via a reconfig.

Read more about metrics in the Metrics section in the developer documentation. Statistics Service

The Statistics Service (stats service for short) supports for collecting arbitrary data from within a running Buildbot instance and export it do a number of storage backends. Currently, only InfluxDB is supported as a storage backend. Also, InfluxDB (or any other storage backend) is not a mandatory dependency. Buildbot can run without it although StatsService will be of no use in such a case. At present, StatsService can keep track of build properties, build times (start, end, duration) and arbitrary data produced inside Buildbot (more on this later).

Example usage:

captures = [stats.CaptureProperty('Builder1', 'tree-size-KiB'),
c['services'] = []
        stats.InfluxStorageService('localhost', 8086, 'root', 'root', 'test', captures)
    ], name="StatsService"))

The services configuration value should be initialized as a list and a StatsService instance should be appended to it as shown in the example above.

Statistics Service
class buildbot.statistics.stats_service.StatsService

This is the main class for statistics service. It is initialized in the master configuration as show in the example above. It takes two arguments:

A list of storage backends (see Storage Backends). In the example above, stats.InfluxStorageService is an instance of a storage backend. Each storage backend is an instances of subclasses of statsStorageBase.
The name of this service.

yieldMetricsValue: This method can be used to send arbitrary data for storage. (See Using StatsService.yieldMetricsValue for more information.)

Capture Classes
class buildbot.statistics.capture.CaptureProperty

Instance of this class declares which properties must be captured and sent to the Storage Backends. It takes the following arguments:

The name of builder in which the property is recorded.
The name of property needed to be recorded as a statistic.
(Optional) A custom callback function for this class. This callback function should take in two arguments - build_properties (dict) and property_name (str) and return a string that will be sent for storage in the storage backends.
If this is set to True, then the property name can be a regular expression. All properties matching this regular expression will be sent for storage.
class buildbot.statistics.capture.CapturePropertyAllBuilders

Instance of this class declares which properties must be captured on all builders and sent to the Storage Backends. It takes the following arguments:

The name of property needed to be recorded as a statistic.
(Optional) A custom callback function for this class. This callback function should take in two arguments - build_properties (dict) and property_name (str) and return a string that will be sent for storage in the storage backends.
If this is set to True, then the property name can be a regular expression. All properties matching this regular expression will be sent for storage.
class buildbot.statistics.capture.CaptureBuildStartTime

Instance of this class declares which builders' start times are to be captured and sent to Storage Backends. It takes the following arguments:

The name of builder whose times are to be recorded.
(Optional) A custom callback function for this class. This callback function should take in a Python datetime object and return a string that will be sent for storage in the storage backends.
class buildbot.statistics.capture.CaptureBuildStartTimeAllBuilders

Instance of this class declares start times of all builders to be captured and sent to Storage Backends. It takes the following arguments:

(Optional) A custom callback function for this class. This callback function should take in a Python datetime object and return a string that will be sent for storage in the storage backends.
class buildbot.statistics.capture.CaptureBuildEndTime

Exactly like CaptureBuildStartTime except it declares the builders whose end time is to be recorded. The arguments are same as CaptureBuildStartTime.

class buildbot.statistics.capture.CaptureBuildEndTimeAllBuilders

Exactly like CaptureBuildStartTimeAllBuilders except it declares all builders' end time to be recorded. The arguments are same as CaptureBuildStartTimeAllBuilders.

class buildbot.statistics.capture.CaptureBuildDuration

Instance of this class declares the builders whose build durations are to be recorded. It takes the following arguments:

The name of builder whose times are to be recorded.
Can be one of three: 'seconds', 'minutes', or 'hours'. This is the units in which the build time will be reported.
(Optional) A custom callback function for this class. This callback function should take in two Python datetime objects - a start_time and an end_time and return a string that will be sent for storage in the storage backends.
class buildbot.statistics.capture.CaptureBuildDurationAllBuilders

Instance of this class declares build durations to be recorded for all builders. It takes the following arguments:

Can be one of three: 'seconds', 'minutes', or 'hours'. This is the units in which the build time will be reported.
(Optional) A custom callback function for this class. This callback function should take in two Python datetime objects - a start_time and an end_time and return a string that will be sent for storage in the storage backends.
class buildbot.statistics.capture.CaptureData

Instance of this capture class is for capturing arbitrary data that is not stored as build-data. Needs to be used in conjunction with yieldMetricsValue (See Using StatsService.yieldMetricsValue). Takes the following arguments:

The name of data to be captured. Same as in yieldMetricsValue.
The name of builder whose times are to be recorded.
The callback function for this class. This callback receives the data sent to yieldMetricsValue as post_data (See Using StatsService.yieldMetricsValue). It must return a string that is to be sent to the storage backends for storage.
class buildbot.statistics.capture.CaptureDataAllBuilders

Instance of this capture class for capturing arbitrary data that is not stored as build-data on all builders. Needs to be used in conjunction with yieldMetricsValue (See Using StatsService.yieldMetricsValue). Takes the following arguments:

The name of data to be captured. Same as in yieldMetricsValue.
The callback function for this class. This callback receives the data sent to yieldMetricsValue as post_data (See Using StatsService.yieldMetricsValue). It must return a string that is to be sent to the storage backends for storage.
Using StatsService.yieldMetricsValue

Advanced users can modify BuildSteps to use StatsService.yieldMetricsValue which will send arbitrary data for storage to the StatsService. It takes the following arguments:

The name of the data being sent or storage.
A dictionary of key value pair that is sent for storage. The keys will act as columns in a database and the value is stored under that column.
The integer build id of the current build. Obtainable in all BuildSteps.

Along with using yieldMetricsValue, the user will also need to use the CaptureData capture class. As an example, we can add the following to a build step:

yieldMetricsValue('test_data_name', {'some_data': 'some_value'}, buildid)

Then, we can add in the master configuration a capture class like this:

captures = [CaptureBuildData('test_data_name', 'Builder1')]

Pass this captures list to a storage backend (as shown in the example at the top of this section) for capturing this data.

Storage Backends

Storage backends are responsible for storing any statistics data sent to them. A storage backend will generally be some sort of a database-server running on a machine. (Note: This machine may be different from the one running BuildMaster)

Currently, only InfluxDB is supported as a storage backend.

class buildbot.statistics.storage_backends.influxdb_client.InfluxStorageService

This class is a Buildbot client to the InfluxDB storage backend. InfluxDB is a distributed, time series database that employs a key-value pair storage system.

It requires the following arguments:

The URL where the service is running.
The port on which the service is listening.
Username of a InfluxDB user.
Password for user.
The name of database to be used.
A list of objects of Capture Classes. This tells which statistics are to be stored in this storage backend.
(Optional) The name of this storage backend. BuildbotNetUsageData

Since buildbot 0.9.0, buildbot has a simple feature which sends usage analysis info to This is very important for buildbot developers to understand how the community is using the tools. This allows to better prioritize issues, and understand what plugins are actually being used. This will also be a tool to decide whether to keep support for very old tools. For example buildbot contains support for the venerable CVS, but we have no information whether it actually works beyond the unit tests. We rely on the community to test and report issues with the old features.

With BuildbotNetUsageData, we can know exactly what combination of plugins are working together, how much people are customizing plugins, what versions of the main dependencies people run.

We take your privacy very seriously.

BuildbotNetUsageData will never send information specific to your Code or Intellectual Property. No repository url, shell command values, host names, ip address or custom class names. If it does, then this is a bug, please report.

We still need to track unique number for installation. This is done via doing a sha1 hash of master's hostname, installation path and fqdn. Using a secure hash means there is no way of knowing hostname, path and fqdn given the hash, but still there is a different hash for each master.

You can see exactly what is sent in the master's twisted.log. Usage data is sent every time the master is started.

BuildbotNetUsageData can be configured with 4 values:

  • c['buildbotNetUsageData'] = None disables the feature

  • c['buildbotNetUsageData'] = 'basic' sends the basic information to buildbot including:

    • versions of buildbot, python and twisted
    • platform information (CPU, OS, distribution, python flavor (i.e CPython vs PyPy))
    • mq and database type (mysql or sqlite?)
    • www plugins usage
    • Plugins usages: This counts the number of time each class of buildbot is used in your configuration. This counts workers, builders, steps, schedulers, change sources. If the plugin is subclassed, then it will be prefixed with a >

    example of basic report (for the metabuildbot):

    'versions': {
        'Python': '2.7.6',
        'Twisted': '15.5.0',
        'Buildbot': '0.9.0rc2-176-g5fa9dbf'
    'platform': {
        'machine': 'x86_64',
        'python_implementation': 'CPython',
        'version': '#140-Ubuntu SMP Mon Jul',
        'distro:': ('Ubuntu', '14.04', 'trusty')
    'db': 'sqlite',
    'mq': 'simple',
    'plugins': {
        'buildbot.schedulers.forcesched.ForceScheduler': 2,
        'buildbot.schedulers.triggerable.Triggerable': 1,
        'buildbot.config.BuilderConfig': 4,
        'buildbot.schedulers.basic.AnyBranchScheduler': 2,
        'buildbot.steps.source.git.Git': 4,
        '>>buildbot.steps.trigger.Trigger': 2,
        '>>>buildbot.worker.base.Worker': 4,
        'buildbot.reporters.irc.IRC': 1,
        '>>>buildbot.process.buildstep.LoggingBuildStep': 2},
    'www_plugins': ['buildbot_travis', 'waterfall_view']
  • c['buildbotNetUsageData'] = 'full' sends the basic information plus additional information:

    • configuration of each builders: how the steps are arranged together. for ex:
        'builders': [
            ['buildbot.steps.source.git.Git', '>>>buildbot.process.buildstep.LoggingBuildStep'],
            ['buildbot.steps.source.git.Git', '>>buildbot.steps.trigger.Trigger'],
            ['buildbot.steps.source.git.Git', '>>>buildbot.process.buildstep.LoggingBuildStep'],
            ['buildbot.steps.source.git.Git', '>>buildbot.steps.trigger.Trigger']]
  • c['buildbotNetUsageData'] = myCustomFunction. You can also specify exactly what to send using a callback.

    The custom function will take the generated data from full report in the form of a dictionary, and return a customized report as a jsonable dictionary. You can use this to filter any information you dont want to disclose. You can use a custom http_proxy environment variable in order to not send any data while developing your callback. Users Options
from buildbot.plugins import util
c['user_managers'] = []

user_managers contains a list of ways to manually manage User Objects within Buildbot (see User Objects). Currently implemented is a commandline tool buildbot user, described at length in user. In the future, a web client will also be able to manage User Objects and their attributes.

As shown above, to enable the buildbot user tool, you must initialize a CommandlineUserManager instance in your master.cfg. CommandlineUserManager instances require the following arguments:

This is the username that will be registered on the PB connection and need to be used when calling buildbot user.
This is the passwd that will be registered on the PB connection and need to be used when calling buildbot user.
The PB connection port must be different than c['protocols']['pb']['port'] and be specified when calling buildbot user Input Validation
import re
c['validation'] = {
    'branch' : re.compile(r'^[\w.+/~-]*$'),
    'revision' : re.compile(r'^[ \w\.\-\/]*$'),
    'property_name' : re.compile(r'^[\w\.\-\/\~:]*$'),
    'property_value' : re.compile(r'^[\w\.\-\/\~:]*$'),

This option configures the validation applied to user inputs of various types. This validation is important since these values are often included in command-line arguments executed on workers. Allowing arbitrary input from untrusted users may raise security concerns.

The keys describe the type of input validated; the values are compiled regular expressions against which the input will be matched. The defaults for each type of input are those given in the example, above. Codebase Generator
all_repositories = {
    r'https://hg/hg/mailsuite/mailclient': 'mailexe',
    r'https://hg/hg/mailsuite/mapilib': 'mapilib',
    r'https://hg/hg/mailsuite/imaplib': 'imaplib',
    r'': 'mailexe',
    r'': 'mapilib',
    r'': 'imaplib',

def codebaseGenerator(chdict):
    return all_repositories[chdict['repository']]

c['codebaseGenerator'] = codebaseGenerator

For any incoming change, the codebase is set to ''. This codebase value is sufficient if all changes come from the same repository (or clones). If changes come from different repositories, extra processing will be needed to determine the codebase for the incoming change. This codebase will then be a logical name for the combination of repository and or branch etc.

The codebaseGenerator accepts a change dictionary as produced by the buildbot.db.changes.ChangesConnectorComponent, with a changeid equal to None.

2.4.3. Change Sources

A Version Control System maintains a source tree, and tells the buildmaster when it changes. The first step of each Build is typically to acquire a copy of some version of this tree.

This chapter describes how the Buildbot learns about what Changes have occurred. For more information on VC systems and Changes, see Version Control Systems.

Changes can be provided by a variety of ChangeSource types, although any given project will typically have only a single ChangeSource active. This section provides a description of all available ChangeSource types and explains how to set up each of them. Choosing a Change Source

There are a variety of ChangeSource classes available, some of which are meant to be used in conjunction with other tools to deliver Change events from the VC repository to the buildmaster.

As a quick guide, here is a list of VC systems and the ChangeSources that might be useful with them. Note that some of these modules are in Buildbot's "contrib" directory, meaning that they have been offered by other users in hopes they may be useful, and might require some additional work to make them functional.

  • CVSMaildirSource (watching mail sent by contrib/ script)
  • PBChangeSource (listening for connections from buildbot sendchange run in a loginfo script)
  • PBChangeSource (listening for connections from a long-running contrib/ polling process which examines the ViewCVS database directly)
  • Change Hooks in WebStatus
Bzr (the newer Bazaar)
  • PBChangeSource (listening for connections from contrib/ run in a post-change-branch-tip or commit hook)
  • BzrPoller (polling the Bzr repository)
  • Change Hooks in WebStatus
  • PBChangeSource (listening for connections from contrib/ run in the post-receive hook)
  • PBChangeSource (listening for connections from contrib/, which listens for notifications from GitHub)
  • Change Hooks in WebStatus
  • GitHub change hook (specifically designed for GitHub notifications, but requiring a publicly-accessible WebStatus)
  • BitBucket change hook (specifically designed for BitBucket notifications, but requiring a publicly-accessible WebStatus)
  • GitPoller (polling a remote Git repository)
  • GoogleCodeAtomPoller (polling the commit feed for a GoogleCode Git repository)
  • BitbucketPullrequestPoller (polling Bitbucket for pull requests)
  • PBChangeSource (listening for connections from monotone-buildbot.lua, which is available with Monotone)

All VC systems can be driven by a PBChangeSource and the buildbot sendchange tool run from some form of commit script. If you write an email parsing function, they can also all be driven by a suitable mail-parsing source. Additionally, handlers for web-based notification (i.e. from GitHub) can be used with WebStatus' change_hook module. The interface is simple, so adding your own handlers (and sharing!) should be a breeze.

See Change Source Index for a full list of change sources. Configuring Change Sources

The change_source configuration key holds all active change sources for the configuration.

Most configurations have a single ChangeSource, watching only a single tree, e.g.,

from buildbot.plugins import changes
c['change_source'] = changes.PBChangeSource()

For more advanced configurations, the parameter can be a list of change sources:

source1 = ...
source2 = ...
c['change_source'] = [
    source1, source1
Repository and Project

ChangeSources will, in general, automatically provide the proper repository attribute for any changes they produce. For systems which operate on URL-like specifiers, this is a repository URL. Other ChangeSources adapt the concept as necessary.

Many ChangeSources allow you to specify a project, as well. This attribute is useful when building from several distinct codebases in the same buildmaster: the project string can serve to differentiate the different codebases. Schedulers can filter on project, so you can configure different builders to run for each project. Mail-parsing ChangeSources

Many projects publish information about changes to their source tree by sending an email message out to a mailing list, frequently named PROJECT-commits or PROJECT-changes. Each message usually contains a description of the change (who made the change, which files were affected) and sometimes a copy of the diff. Humans can subscribe to this list to stay informed about what's happening to the source tree.

The Buildbot can also be subscribed to a -commits mailing list, and can trigger builds in response to Changes that it hears about. The buildmaster admin needs to arrange for these email messages to arrive in a place where the buildmaster can find them, and configure the buildmaster to parse the messages correctly. Once that is in place, the email parser will create Change objects and deliver them to the schedulers (see Schedulers) just like any other ChangeSource.

There are two components to setting up an email-based ChangeSource. The first is to route the email messages to the buildmaster, which is done by dropping them into a maildir. The second is to actually parse the messages, which is highly dependent upon the tool that was used to create them. Each VC system has a collection of favorite change-emailing tools, and each has a slightly different format, so each has a different parsing function. There is a separate ChangeSource variant for each parsing function.

Once you've chosen a maildir location and a parsing function, create the change source and put it in change_source:

from buildbot.plugins import changes
c['change_source'] = changes.CVSMaildirSource("~/maildir-buildbot",
Subscribing the Buildmaster

The recommended way to install the buildbot is to create a dedicated account for the buildmaster. If you do this, the account will probably have a distinct email address (perhaps Then just arrange for this account's email to be delivered to a suitable maildir (described in the next section).

If the buildbot does not have its own account, extension addresses can be used to distinguish between email intended for the buildmaster and email intended for the rest of the account. In most modern MTAs, the e.g. account has control over every email address at which begins with "foo", such that email addressed to can be delivered to a different destination than qmail does this by using separate .qmail files for the two destinations (.qmail-foo and .qmail-bar, with .qmail controlling the base address and .qmail-default controlling all other extensions). Other MTAs have similar mechanisms.

Thus you can assign an extension address like to the buildmaster, and retain for your own use.

Using Maildirs

A maildir is a simple directory structure originally developed for qmail that allows safe atomic update without locking. Create a base directory with three subdirectories: new, tmp, and cur. When messages arrive, they are put into a uniquely-named file (using pids, timestamps, and random numbers) in tmp. When the file is complete, it is atomically renamed into new. Eventually the buildmaster notices the file in new, reads and parses the contents, then moves it into cur. A cronjob can be used to delete files in cur at leisure.

Maildirs are frequently created with the maildirmake tool, but a simple mkdir -p ~/MAILDIR/cur,new,tmp is pretty much equivalent.

Many modern MTAs can deliver directly to maildirs. The usual .forward or .procmailrc syntax is to name the base directory with a trailing slash, so something like ~/MAILDIR/. qmail and postfix are maildir-capable MTAs, and procmail is a maildir-capable MDA (Mail Delivery Agent).

Here is an example procmail config, located in ~/.procmailrc:

# .procmailrc
# routes incoming mail to appropriate mailboxes


If procmail is not setup on a system wide basis, then the following one-line .forward file will invoke it.


For MTAs which cannot put files into maildirs directly, the safecat tool can be executed from a .forward file to accomplish the same thing.

The Buildmaster uses the linux DNotify facility to receive immediate notification when the maildir's new directory has changed. When this facility is not available, it polls the directory for new messages, every 10 seconds by default.

Parsing Email Change Messages

The second component to setting up an email-based ChangeSource is to parse the actual notices. This is highly dependent upon the VC system and commit script in use.

A couple of common tools used to create these change emails, along with the buildbot tools to parse them, are:

Buildbot CVS MailNotifier

The following sections describe the parsers available for each of these tools.

Most of these parsers accept a prefix= argument, which is used to limit the set of files that the buildmaster pays attention to. This is most useful for systems like CVS and SVN which put multiple projects in a single repository (or use repository names to indicate branches). Each filename that appears in the email is tested against the prefix: if the filename does not start with the prefix, the file is ignored. If the filename does start with the prefix, that prefix is stripped from the filename before any further processing is done. Thus the prefix usually ends with a slash.

class buildbot.changes.mail.CVSMaildirSource

This parser works with the script in the contrib directory.

The script sends an email containing all the files submitted in one directory. It is invoked by using the CVSROOT/loginfo facility.

The Buildbot's CVSMaildirSource knows how to parse these messages and turn them into Change objects. It takes the directory name of the maildir root. For example:

from buildbot.plugins import changes
c['change_source'] = changes.CVSMaildirSource("/home/buildbot/Mail")
Configuration of CVS and

CVS must be configured to invoke the script when files are checked in. This is done via the CVS loginfo configuration file.

To update this, first do:

cvs checkout CVSROOT

cd to the CVSROOT directory and edit the file loginfo, adding a line like:

SomeModule /cvsroot/CVSROOT/ --cvsroot -e buildbot -P SomeModule %@{sVv@}


For cvs version 1.12.x, the --path %p option is required. Version 1.11.x and 1.12.x report the directory path differently.

The above example you put the script under /cvsroot/CVSROOT. It can be anywhere. Run the script with --help to see all the options. At the very least, the options -e (email) and -P (project) should be specified. The line must end with %{sVv}. This is expanded to the files that were modified.

Additional entries can be added to support more modules.

See --help` for more information on the available options.

class buildbot.changes.mail.SVNCommitEmailMaildirSource

SVNCommitEmailMaildirSource parses message sent out by the script, which is included in the Subversion distribution.

It does not currently handle branches: all of the Change objects that it creates will be associated with the default (i.e. trunk) branch.

from buildbot.plugins import changes
c['change_source'] = changes.SVNCommitEmailMaildirSource("~/maildir-buildbot")
class buildbot.changes.mail.BzrLaunchpadEmailMaildirSource

BzrLaunchpadEmailMaildirSource parses the mails that are sent to addresses that subscribe to branch revision notifications for a bzr branch hosted on Launchpad.

The branch name defaults to lp:Launchpad path. For example lp:~maria-captains/maria/5.1.

If only a single branch is used, the default branch name can be changed by setting defaultBranch.

For multiple branches, pass a dictionary as the value of the branchMap option to map specific repository paths to specific branch names (see example below). The leading lp: prefix of the path is optional.

The prefix option is not supported (it is silently ignored). Use the branchMap and defaultBranch instead to assign changes to branches (and just do not subscribe the buildbot to branches that are not of interest).

The revision number is obtained from the email text. The bzr revision id is not available in the mails sent by Launchpad. However, it is possible to set the bzr append_revisions_only option for public shared repositories to avoid new pushes of merges changing the meaning of old revision numbers.

from buildbot.plugins import changes
bm = {
    'lp:~maria-captains/maria/5.1': '5.1',
    'lp:~maria-captains/maria/6.0': '6.0'
c['change_source'] = changes.BzrLaunchpadEmailMaildirSource("~/maildir-buildbot",
                                                            branchMap=bm) PBChangeSource
class buildbot.changes.pb.PBChangeSource

PBChangeSource actually listens on a TCP port for clients to connect and push change notices into the Buildmaster. This is used by the built-in buildbot sendchange notification tool, as well as several version-control hook scripts. This change is also useful for creating new kinds of change sources that work on a push model instead of some kind of subscription scheme, for example a script which is run out of an email .forward file. This ChangeSource always runs on the same TCP port as the workers. It shares the same protocol, and in fact shares the same space of "usernames", so you cannot configure a PBChangeSource with the same name as a worker.

If you have a publicly accessible worker port, and are using PBChangeSource, you must establish a secure username and password for the change source. If your sendchange credentials are known (e.g., the defaults), then your buildmaster is susceptible to injection of arbitrary changes, which (depending on the build factories) could lead to arbitrary code execution on workers.

The PBChangeSource is created with the following arguments.

which port to listen on. If None (which is the default), it shares the port used for worker connections.
The user account that the client program must use to connect. Defaults to change
The password for the connection - defaults to changepw. Do not use this default on a publicly exposed port!

The prefix to be found and stripped from filenames delivered over the connection, defaulting to None. Any filenames which do not start with this prefix will be removed. If all the filenames in a given Change are removed, the that whole Change will be dropped. This string should probably end with a directory separator.

This is useful for changes coming from version control systems that represent branches as parent directories within the repository (like SVN and Perforce). Use a prefix of trunk/ or project/branches/foobranch/ to only follow one branch and to get correct tree-relative filenames. Without a prefix, the PBChangeSource will probably deliver Changes with filenames like trunk/foo.c instead of just foo.c. Of course this also depends upon the tool sending the Changes in (like buildbot sendchange) and what filenames it is delivering: that tool may be filtering and stripping prefixes at the sending end.

For example:

from buildbot.plugins import changes
c['change_source'] = changes.PBChangeSource(port=9999, user='laura', passwd='fpga')

The following hooks are useful for sending changes to a PBChangeSource:

Bzr Hook

Bzr is also written in Python, and the Bzr hook depends on Twisted to send the changes.

To install, put contrib/ in one of your plugins locations a bzr plugins directory (e.g., ~/.bazaar/plugins). Then, in one of your bazaar conf files (e.g., ~/.bazaar/locations.conf), set the location you want to connect with buildbot with these keys:

  • buildbot_on one of 'commit', 'push, or 'change'. Turns the plugin on to report changes via commit, changes via push, or any changes to the trunk. 'change' is recommended.
  • buildbot_server (required to send to a buildbot master) the URL of the buildbot master to which you will connect (as of this writing, the same server and port to which workers connect).
  • buildbot_port (optional, defaults to 9989) the port of the buildbot master to which you will connect (as of this writing, the same server and port to which workers connect)
  • buildbot_pqm (optional, defaults to not pqm) Normally, the user that commits the revision is the user that is responsible for the change. When run in a pqm (Patch Queue Manager, see environment, the user that commits is the Patch Queue Manager, and the user that committed the parent revision is responsible for the change. To turn on the pqm mode, set this value to any of (case-insensitive) "Yes", "Y", "True", or "T".
  • buildbot_dry_run (optional, defaults to not a dry run) Normally, the post-commit hook will attempt to communicate with the configured buildbot server and port. If this parameter is included and any of (case-insensitive) "Yes", "Y", "True", or "T", then the hook will simply print what it would have sent, but not attempt to contact the buildbot master.
  • buildbot_send_branch_name (optional, defaults to not sending the branch name) If your buildbot's bzr source build step uses a repourl, do not turn this on. If your buildbot's bzr build step uses a baseURL, then you may set this value to any of (case-insensitive) "Yes", "Y", "True", or "T" to have the buildbot master append the branch name to the baseURL.


The bzr smart server (as of version 2.2.2) doesn't know how to resolve bzr:// urls into absolute paths so any paths in locations.conf won't match, hence no change notifications will be sent to Buildbot. Setting configuration parameters globally or in-branch might still work. When buildbot no longer has a hardcoded password, it will be a configuration option here as well.

Here's a simple example that you might have in your ~/.bazaar/locations.conf.

buildbot_on = change
buildbot_server = localhost P4Source

The P4Source periodically polls a Perforce depot for changes. It accepts the following arguments:

The Perforce server to connect to (as host:port).
The Perforce user.
The Perforce password.
The base depot path to watch, without the trailing '/...'.
An optional string parameter. Specify the location of the perforce command line binary (p4). You only need to do this if the perforce binary is not in the path of the buildbot user. Defaults to p4.
A function that maps a pathname, without the leading p4base, to a (branch, filename) tuple. The default just returns (None, branchfile), which effectively disables branch support. You should supply a function which understands your repository structure.
How often to poll, in seconds. Defaults to 600 (10 minutes).
Set the name of the project to be used for the P4Source. This will then be set in any changes generated by the P4Source, and can be used in a Change Filter for triggering particular builders.
Determines when the first poll occurs. True = immediately on launch, False = wait for one pollInterval (default).
The maximum number of changes to inspect at a time. If more than this number occur since the last poll, older changes will be silently ignored.
The character encoding of p4's output. This defaults to "utf8", but if your commit messages are in another encoding, specify that here. For example, if you're using Perforce on Windows, you may need to use "cp437" as the encoding if "utf8" generates errors in your master log.
The timezone of the Perforce server, using the usual timezone format (e.g: "Europe/Stockholm") in case it's not in UTC.
Set to True to use ticket-based authentication, instead of passwords (but you still need to specify p4passwd).
How often to get a new ticket, in seconds, when use_tickets is enabled. Defaults to 86400 (24 hours).

This configuration uses the P4PORT, P4USER, and P4PASSWD specified in the buildmaster's environment. It watches a project in which the branch name is simply the next path component, and the file is all path components after.

from buildbot.plugins import changes
s = changes.P4Source(p4base='//depot/project/',
                     split_file=lambda branchfile: branchfile.split('/',1))
c['change_source'] = s SVNPoller
class buildbot.changes.svnpoller.SVNPoller

The SVNPoller is a ChangeSource which periodically polls a Subversion repository for new revisions, by running the svn log command in a subshell. It can watch a single branch or multiple branches.

SVNPoller accepts the following arguments:


The base URL path to watch, like svn://, or, or even file:///home/svn/Repository/ProjectA/branches/1.5/. This must include the access scheme, the location of the repository (both the hostname for remote ones, and any additional directory names necessary to get to the repository), and the sub-path within the repository's virtual filesystem for the project and branch of interest.

The SVNPoller will only pay attention to files inside the subdirectory specified by the complete repourl.


A function to convert pathnames into (branch, relative_pathname) tuples. Use this to explain your repository's branch-naming policy to SVNPoller. This function must accept a single string (the pathname relative to the repository) and return a two-entry tuple. Directory pathnames always end with a right slash to distinguish them from files, like trunk/src/, or src/. There are a few utility functions in buildbot.changes.svnpoller that can be used as a split_file function; see below for details.

For directories, the relative pathname returned by split_file should end with a right slash but an empty string is also accepted for the root, like ("branches/1.5.x", "") being converted from "branches/1.5.x/".

The default value always returns (None, path), which indicates that all files are on the trunk.

Subclasses of SVNPoller can override the split_file method instead of using the split_file= argument.

Set the name of the project to be used for the SVNPoller. This will then be set in any changes generated by the SVNPoller, and can be used in a Change Filter for triggering particular builders.
An optional string parameter. If set, the option --user argument will be added to all svn commands. Use this if you have to authenticate to the svn server before you can do svn info or svn log commands.
Like svnuser, this will cause a option --password argument to be passed to all svn commands.
How often to poll, in seconds. Defaults to 600 (checking once every 10 minutes). Lower this if you want the buildbot to notice changes faster, raise it if you want to reduce the network and CPU load on your svn server. Please be considerate of public SVN repositories by using a large interval when polling them.
Determines when the first poll occurs. True = immediately on launch, False = wait for one pollInterval (default).
The maximum number of changes to inspect at a time. Every pollInterval seconds, the SVNPoller asks for the last histmax changes and looks through them for any revisions it does not already know about. If more than histmax revisions have been committed since the last poll, older changes will be silently ignored. Larger values of histmax will cause more time and memory to be consumed on each poll attempt. histmax defaults to 100.
This controls the svn executable to use. If subversion is installed in a weird place on your system (outside of the buildmaster's PATH), use this to tell SVNPoller where to find it. The default value of svn will almost always be sufficient.
This parameter is deprecated in favour of specifying a global revlink option. This parameter allows a link to be provided for each revision (for example, to websvn or viewvc). These links appear anywhere changes are shown, such as on build or change pages. The proper form for this parameter is an URL with the portion that will substitute for a revision number replaced by ''%s''. For example, 'http://myserver/websvn/revision.php?rev=%s' could be used to cause revision links to be created to a websvn repository viewer.
If specified, this is a pathname of a cache file that SVNPoller will use to store its state between restarts of the master.
If specified, the extra arguments will be added to the svn command args.

Several split file functions are available for common SVN repository layouts. For a poller that is only monitoring trunk, the default split file function is available explicitly as split_file_alwaystrunk:

from buildbot.plugins import changes, util
c['change_source'] = changes.SVNPoller(

For repositories with the /trunk and /branches/BRANCH layout, split_file_branches will do the job:

from buildbot.plugins import changes, util
c['change_source'] = changes.SVNPoller(

When using this splitter the poller will set the project attribute of any changes to the project attribute of the poller.

For repositories with the PROJECT/trunk and PROJECT/branches/BRANCH layout, split_file_projects_branches will do the job:

from buildbot.plugins import changes, util
c['change_source'] = changes.SVNPoller(

When using this splitter the poller will set the project attribute of any changes to the project determined by the splitter.

The SVNPoller is highly adaptable to various Subversion layouts. See Customizing SVNPoller for details and some common scenarios. Bzr Poller

If you cannot insert a Bzr hook in the server, you can use the Bzr Poller. To use, put contrib/ somewhere that your buildbot configuration can import it. Even putting it in the same directory as the master.cfg should work. Install the poller in the buildbot configuration as with any other change source. Minimally, provide a URL that you want to poll (bzr://, bzr+ssh://, or lp:), making sure the buildbot user has necessary privileges.

# put file to the same directory as master.cfg
from bzr_buildbot import BzrPoller
c['change_source'] = BzrPoller(

The BzrPoller parameters are:

The URL to poll.
The number of seconds to wait between polls. Defaults to 10 minutes.
Any value to be used as the branch name. Defaults to None, or specify a string, or specify the constants from SHORT or FULL to get the short branch name or full branch address.
normally, the user that commits the revision is the user that is responsible for the change. When run in a pqm (Patch Queue Manager, see environment, the user that commits is the Patch Queue Manager, and the user that committed the merged, parent revision is responsible for the change. Set this value to True if this is pointed against a PQM-managed branch. GitPoller

If you cannot take advantage of post-receive hooks as provided by contrib/ for example, then you can use the GitPoller.

The GitPoller periodically fetches from a remote Git repository and processes any changes. It requires its own working directory for operation. The default should be adequate, but it can be overridden via the workdir property.


There can only be a single GitPoller pointed at any given repository.

The GitPoller requires Git-1.7 and later. It accepts the following arguments:

the git-url that describes the remote repository, e.g. (see the git fetch help for more info on git-url formats)

One of the following:

  • a list of the branches to fetch.
  • True indicating that all branches should be fetched
  • a callable which takes a single argument. It should take a remote refspec (such as 'refs/heads/master', and return a boolean indicating whether that branch should be fetched.
accepts a single branch name to fetch. Exists for backwards compatibility with old configurations.
interval in seconds between polls, default is 10 minutes
Determines when the first poll occurs. True = immediately on launch, False = wait for one pollInterval (default).
Determine if a push on a new branch with already known commits should trigger a build. (defaults to False).
path to the Git binary, defaults to just 'git'
Set the category to be used for the changes produced by the GitPoller. This will then be set in any changes generated by the GitPoller, and can be used in a Change Filter for triggering particular builders.
Set the name of the project to be used for the GitPoller. This will then be set in any changes generated by the GitPoller, and can be used in a Change Filter for triggering particular builders.
parse each revision's commit timestamp (default is True), or ignore it in favor of the current time (so recently processed commits appear together in the waterfall page)
Set encoding will be used to parse author's name and commit message. Default encoding is 'utf-8'. This will not be applied to file names since Git will translate non-ascii file names to unreadable escape sequences.
the directory where the poller should keep its local repository. The default is gitpoller_work. If this is a relative path, it will be interpreted relative to the master's basedir. Multiple Git pollers can share the same directory.

A configuration for the Git poller might look like this:

from buildbot.plugins import changes
c['change_source'] = changes.GitPoller(repourl='',
                                       branches=['master', 'great_new_feature']) HgPoller

The HgPoller periodically pulls a named branch from a remote Mercurial repository and processes any changes. It requires its own working directory for operation, which must be specified via the workdir property.

The HgPoller requires a working hg executable, and at least a read-only access to the repository it polls (possibly through ssh keys or by tweaking the hgrc of the system user buildbot runs as).

The HgPoller will not transmit any change if there are several heads on the watched named branch. This is similar (although not identical) to the Mercurial executable behaviour. This exceptional condition is usually the result of a developer mistake, and usually does not last for long. It is reported in logs. If fixed by a later merge, the buildmaster administrator does not have anything to do: that merge will be transmitted, together with the intermediate ones.

The HgPoller accepts the following arguments:

the name of the poller. This must be unique, and defaults to the repourl.
the url that describes the remote repository, e.g. Any url suitable for hg pull can be specified.
the desired branch to pull, will default to 'default'

the directory where the poller should keep its local repository. It is mandatory for now, although later releases may provide a meaningful default.

It also serves to identify the poller in the buildmaster internal database. Changing it may result in re-processing all changes so far.

Several HgPoller instances may share the same workdir for mutualisation of the common history between two different branches, thus easing on local and remote system resources and bandwidth.

If relative, the workdir will be interpreted from the master directory.

interval in seconds between polls, default is 10 minutes
Determines when the first poll occurs. True = immediately on launch, False = wait for one pollInterval (default).
path to the Mercurial binary, defaults to just 'hg'
Set the category to be used for the changes produced by the HgPoller. This will then be set in any changes generated by the HgPoller, and can be used in a Change Filter for triggering particular builders.
Set the name of the project to be used for the HgPoller. This will then be set in any changes generated by the HgPoller, and can be used in a Change Filter for triggering particular builders.
parse each revision's commit timestamp (default is True), or ignore it in favor of the current time (so recently processed commits appear together in the waterfall page)
Set encoding will be used to parse author's name and commit message. Default encoding is 'utf-8'.

A configuration for the Mercurial poller might look like this:

from buildbot.plugins import changes
c['change_source'] = changes.HgPoller(repourl='',
                                      workdir='hg-myrepo') BitbucketPullrequestPoller
class buildbot.changes.bitbucket.BitbucketPullrequestPoller

This BitbucketPullrequestPoller periodically polls Bitbucket for new or updated pull requests. It uses Bitbuckets powerful Pull Request REST API to gather the information needed.

The BitbucketPullrequestPoller accepts the following arguments:

The owner of the Bitbucket repository. All Bitbucket Urls are of the form
The name of the Bitbucket repository.
A single branch or a list of branches which should be processed. If it is None (the default) all pull requests are used.
Interval in seconds between polls, default is 10 minutes.
Determines when the first poll occurs. True = immediately on launch, False = wait for one pollInterval (default).
Set the category to be used for the changes produced by the BitbucketPullrequestPoller. This will then be set in any changes generated by the BitbucketPullrequestPoller, and can be used in a Change Filter for triggering particular builders.
Set the name of the project to be used for the BitbucketPullrequestPoller. This will then be set in any changes generated by the BitbucketPullrequestPoller, and can be used in a Change Filter for triggering particular builders.
A callable which takes one parameter, the decoded Python object of the pull request JSON. If the it returns False the pull request is ignored. It can be used to define custom filters based on the content of the pull request. See the Bitbucket documentation for more information about the format of the response. By default the filter always returns True.
parse each revision's commit timestamp (default is True), or ignore it in favor of the current time (so recently processed commits appear together in the waterfall page)
Set encoding will be used to parse author's name and commit message. Default encoding is 'utf-8'.

A minimal configuration for the Bitbucket pull request poller might look like this:

from buildbot.plugins import changes
c['change_source'] = changes.BitbucketPullrequestPoller(

Here is a more complex configuration using a pullrequest_filter. The pull request is only processed if at least 3 people have already approved it:

def approve_filter(pr, threshold):
    approves = 0
    for participant in pr['participants']:
        if participant['approved']:
            approves = approves + 1

    if approves < threshold:
        return False
    return True

from buildbot.plugins import changes
c['change_source'] = changes.BitbucketPullrequestPoller(
    pullrequest_filter=lambda pr : approve_filter(pr,3),


Anyone who can create pull requests for the Bitbucket repository can initiate a change, potentially causing the buildmaster to run arbitrary code. GerritChangeSource
class buildbot.changes.gerritchangesource.GerritChangeSource

The GerritChangeSource class connects to a Gerrit server by its SSH interface and uses its event source mechanism, gerrit stream-events.

The GerritChangeSource accepts the following arguments:

the dns or ip that host the gerrit ssh server
the port of the gerrit ssh server
the username to use to connect to gerrit
ssh identity file to for authentication (optional). Pay attention to the ssh passphrase
event to be handled (optional). By default processes patchset-created and ref-updated
Print gerrit event in the log (default False). This allows to debug event content, but will eventually fill your logs with useless gerrit event logs.

By default this class adds a change to the buildbot system for each of the following events:

A change is proposed for review. Automatic checks like can be automatically triggered. Beware of what kind of automatic task you trigger. At this point, no trusted human has reviewed the code, and a patch could be specially crafted by an attacker to compromise your workers.
A change has been merged into the repository. Typically, this kind of event can lead to a complete rebuild of the project, and upload binaries to an incremental build results server.

But you can specify how to handle Events:

  • Any event with change and patchSet will be processed by universal collector by default.
  • In case you've specified processing function for the given kind of events, all events of this kind will be processed only by this function, bypassing universal collector.

An example:

from buildbot.plugins import changes
class MyGerritChangeSource(changes.GerritChangeSource):
    """Custom GerritChangeSource
    def eventReceived_patchset_created(self, properties, event):
        """Handler events without properties
        properties = {}
        self.addChangeFromEvent(properties, event)

This class will populate the property list of the triggered build with the info received from Gerrit server in JSON format.

In case of patchset-created event, these properties will be:

Branch of the Change
Change's ID in the Gerrit system (the ChangeId: in commit comments)
Change's number in Gerrit system
Change's owner email (owner is first uploader)
Change's owner name
Project of the Change
Change's subject
URL of the Change in the Gerrit's web interface
Patchset's version number
Patchset's Gerrit "virtual branch"
Patchset's Git commit ID
Patchset uploader's email (owner is first uploader)
Patchset uploader's name (owner is first uploader)
Event type (patchset-created)
Patchset uploader's email
Patchset uploader's name

In case of ref-updated event, these properties will be:

New Git commit ID (after merger)
Previous Git commit ID (before merger)
Project that was updated
Branch that was updated
Submitter's email (merger responsible)
Submitter's name (merger responsible)
Event type (ref-updated)
Submitter's email (merger responsible)
Submitter's name (merger responsible)

A configuration for this source might look like:

from buildbot.plugins import changes
c['change_source'] = changes.GerritChangeSource(
    handled_events=["patchset-created", "change-merged"])

see master/docs/examples/git_gerrit.cfg or master/docs/examples/repo_gerrit.cfg in the Buildbot distribution for a full example setup of Git+Gerrit or Repo+Gerrit of GerritChangeSource. GerritChangeFilter
class buildbot.changes.gerritchangesource.GerritChangeFilter

GerritChangeFilter is a ready to use ChangeFilter you can pass to AnyBranchScheduler in order to filter changes, to create pre-commit builders or post-commit schedulers. It has the same api as Change Filter, except it has additionnal eventtype set of filter (can as well be specified as value, list, regular expression or callable)

An example is following:

from buildbot.plugins import schedulers, util
# this scheduler will create builds when a patch is uploaded to gerrit
# but only if it is uploaded to the "main" branch

# this scheduler will create builds when a patch is merged in the "main" branch
# for post-commit tests
                              change_filter=util.GerritChangeFilter("main", "ref-updated"),
                              builderNames=["main-postcommit"]) Change Hooks (HTTP Notifications)

Buildbot already provides a web frontend, and that frontend can easily be used to receive HTTP push notifications of commits from services like GitHub or GoogleCode. See Change Hooks for more information. GoogleCodeAtomPoller

The GoogleCodeAtomPoller periodically polls a Google Code Project's commit feed for changes. Works on SVN, Git, and Mercurial repositories. Branches are not understood (yet). It accepts the following arguments:

The commit Atom feed URL of the GoogleCode repository (MANDATORY)
Polling frequency for the feed (in seconds). Default is 1 hour (OPTIONAL)

As an example, to poll the Ostinato project's commit feed every 3 hours, the configuration would look like this:

from googlecode_atom import GoogleCodeAtomPoller
c['change_source'] = GoogleCodeAtomPoller(


You will need to download from the Buildbot source and install it somewhere on your PYTHONPATH first.

2.4.4. Schedulers

Schedulers are responsible for initiating builds on builders.

Some schedulers listen for changes from ChangeSources and generate build sets in response to these changes. Others generate build sets without changes, based on other events in the buildmaster. Configuring Schedulers

The schedulers configuration parameter gives a list of scheduler instances, each of which causes builds to be started on a particular set of Builders. The two basic scheduler classes you are likely to start with are SingleBranchScheduler and Periodic, but you can write a customized subclass to implement more complicated build scheduling.

Scheduler arguments should always be specified by name (as keyword arguments), to allow for future expansion:

sched = SingleBranchScheduler(name="quick", builderNames=['lin', 'win'])

There are several common arguments for schedulers, although not all are available with all schedulers.

Each Scheduler must have a unique name. This is used in status displays, and is also available in the build property scheduler.
This is the set of builders which this scheduler should trigger, specified as a list of names (strings).

This is a dictionary specifying properties that will be transmitted to all builds started by this scheduler. The owner property may be of particular interest, as its contents (as a list) will be added to the list of "interested users" (Doing Things With Users) for each triggered build. For example

sched = Scheduler(...,
    properties = {
        'owner': ['', '']
A callable which takes one argument, a Change instance, and returns True if the change is worth building, and False if it is not. Unimportant Changes are accumulated until the build is triggered by an important change. The default value of None means that all Changes are important.
The change filter that will determine which changes are recognized by this scheduler; Change Filters. Note that this is different from fileIsImportant: if the change filter filters out a Change, then it is completely ignored by the scheduler. If a Change is allowed by the change filter, but is deemed unimportant, then it will not cause builds to start, but will be remembered and shown in status displays.

When the scheduler processes data from more than one repository at the same time, a corresponding codebase definition should be passed for each repository.

This parameter can be specified either as a list of strings (simplest form; use if no special overrides are needed) or as a dictionary of dictionaries (where each dict is a codebase definition as described next).

Each codebase definition is a dictionary with any of the keys: repository, branch, revision. The codebase definitions are combined in a dictionary keyed by the name of the codebase.

codebases = {'codebase1': {'repository':'....',
                           'revision': None},
             'codebase2': {'repository':'....'} }


The codebases parameter is only used to fill in missing details about a codebases when scheduling a build. For example, when a change to codebase A occurs, a scheduler must invent a sourcestamp for codebase B. The parameter does not act as a filter on incoming changes -- use a change filter for that purpose.

Source steps can specify a codebase to which they will apply, and will use the sourcestamp for that codebase.

A boolean that, when True, only adds important changes to the buildset as specified in the fileIsImportant callable. This means that unimportant changes are ignored the same way a change_filter filters changes. This defaults to False and only applies when fileIsImportant is given.
A string that will be used as the reason for the triggered build.

The remaining subsections represent a catalog of the available scheduler types. All these schedulers are defined in modules under buildbot.schedulers, and the docstrings there are the best source of documentation on the arguments taken by each one. Scheduler Resiliency

In a multi-master configuration, schedulers with the same name can be configured on multiple masters. Only one instance of the scheduler will be active. If that instance becomes inactive, due to its master being shut down or failing, then another instance will become active after a short delay. This provides resiliency in scheduler configurations, so that schedulers are not a single point of failure in a Buildbot infrastructure.

The Data API and web UI display the master on which each scheduler is running.

There is currently no mechanism to control which master's scheduler instance becomes active. The behavior is nondeterministic, based on the timing of polling by inactive schedulers. The failover is non-revertive. Change Filters

Several schedulers perform filtering on an incoming set of changes. The filter can most generically be specified as a ChangeFilter. Set up a ChangeFilter like this:

from buildbot.plugins import util
my_filter = util.ChangeFilter(project_re="^baseproduct/.*", branch="devel")

and then add it to a scheduler with the change_filter parameter:

sch = SomeSchedulerClass(...,

There are five attributes of changes on which you can filter:

the project string, as defined by the ChangeSource.
the repository in which this change occurred.
the branch on which this change occurred. Note that 'trunk' or 'master' is often denoted by None.
the category, again as defined by the ChangeSource.
the change's codebase.

For each attribute, the filter can look for a single, specific value:

my_filter = util.ChangeFilter(project='myproject')

or accept any of a set of values:

my_filter = util.ChangeFilter(project=['myproject', 'jimsproject'])

or apply a regular expression, using the attribute name with a "_re" suffix:

my_filter = util.ChangeFilter(category_re='.*deve.*')
# or, to use regular expression flags:
import re
my_filter = util.ChangeFilter(category_re=re.compile('.*deve.*', re.I))

For anything more complicated, define a Python function to recognize the strings you want:

def my_branch_fn(branch):
    return branch in branches_to_build and branch not in branches_to_ignore
my_filter = util.ChangeFilter(branch_fn=my_branch_fn)

The special argument filter_fn can be used to specify a function that is given the entire Change object, and returns a boolean.

The entire set of allowed arguments, then, is

project project_re project_fn
repository repository_re repository_fn
branch branch_re branch_fn
category category_re category_fn
codebase codebase_re codebase_fn

A Change passes the filter only if all arguments are satisfied. If no filter object is given to a scheduler, then all changes will be built (subject to any other restrictions the scheduler enforces). Scheduler Types

The remaining subsections represent a catalog of the available Scheduler types. All these Schedulers are defined in modules under buildbot.schedulers, and the docstrings there are the best source of documentation on the arguments taken by each one.


This is the original and still most popular scheduler class. It follows exactly one branch, and starts a configurable tree-stable-timer after each change on that branch. When the timer expires, it starts a build on some set of Builders. This scheduler accepts a fileIsImportant function which can be used to ignore some Changes if they do not affect any important files.

If treeStableTimer is not set, then this scheduler starts a build for every Change that matches its change_filter and statsfies fileIsImportant. If treeStableTimer is set, then a build is triggered for each set of Changes which arrive within the configured time, and match the filters.


The behavior of this scheduler is undefined, if treeStableTimer is set, and changes from multiple branches, repositories or codebases are accepted by the filter.


The codebases argument will filter out codebases not specified there, but won't filter based on the branches specified there.

The arguments to this scheduler are:









The scheduler will wait for this many seconds before starting the build. If new changes are made during this interval, the timer will be restarted, so really the build will be started after a change and then after this many seconds of inactivity.

If treeStableTimer is None, then a separate build is started immediately for each Change.

A callable which takes one argument, a Change instance, and returns True if the change is worth building, and False if it is not. Unimportant Changes are accumulated until the build is triggered by an important change. The default value of None means that all Changes are important.
categories (deprecated; use change_filter)
A list of categories of changes that this scheduler will respond to. If this is specified, then any non-matching changes are ignored.
branch (deprecated; use change_filter)

The scheduler will pay attention to this branch, ignoring Changes that occur on other branches. Setting branch equal to the special value of None means it should only pay attention to the default branch.


None is a keyword, not a string, so write None and not "None".


from buildbot.plugins import schedulers, util
quick = schedulers.SingleBranchScheduler(
            builderNames=["quick-linux", "quick-netbsd"])
full = schedulers.SingleBranchScheduler(
            builderNames=["full-linux", "full-netbsd", "full-OSX"])
c['schedulers'] = [quick, full]

In this example, the two quick builders are triggered 60 seconds after the tree has been changed. The full builds do not run quite so quickly (they wait 5 minutes), so hopefully if the quick builds fail due to a missing file or really simple typo, the developer can discover and fix the problem before the full builds are started. Both schedulers only pay attention to the default branch: any changes on other branches are ignored. Each scheduler triggers a different set of Builders, referenced by name.


The old names for this scheduler, buildbot.scheduler.Scheduler and buildbot.schedulers.basic.Scheduler, are deprecated in favor of using buildbot.plugins:

from buildbot.plugins import schedulers

However if you must use a fully qualified name, it is buildbot.schedulers.basic.SingleBranchScheduler.


This scheduler uses a tree-stable-timer like the default one, but uses a separate timer for each branch.

If treeStableTimer is not set, then this scheduler is indistinguishable from bb:sched:SingleBranchScheduler. If treeStableTimer is set, then a build is triggered for each set of Changes which arrive within the configured time, and match the filters.

The arguments to this scheduler are:







See Configuring Schedulers.
The scheduler will wait for this many seconds before starting the build. If new changes are made on the same branch during this interval, the timer will be restarted.
branches (deprecated; use change_filter)
Changes on branches not specified on this list will be ignored.
categories (deprecated; use change_filter)
A list of categories of changes that this scheduler will respond to. If this is specified, then any non-matching changes are ignored.
Dependent Scheduler

It is common to wind up with one kind of build which should only be performed if the same source code was successfully handled by some other kind of build first. An example might be a packaging step: you might only want to produce .deb or RPM packages from a tree that was known to compile successfully and pass all unit tests. You could put the packaging step in the same Build as the compile and testing steps, but there might be other reasons to not do this (in particular you might have several Builders worth of compiles/tests, but only wish to do the packaging once). Another example is if you want to skip the full builds after a failing quick build of the same source code. Or, if one Build creates a product (like a compiled library) that is used by some other Builder, you'd want to make sure the consuming Build is run after the producing one.

You can use dependencies to express this relationship to the Buildbot. There is a special kind of scheduler named Dependent that will watch an upstream scheduler for builds to complete successfully (on all of its Builders). Each time that happens, the same source code (i.e. the same SourceStamp) will be used to start a new set of builds, on a different set of Builders. This downstream scheduler doesn't pay attention to Changes at all. It only pays attention to the upstream scheduler.

If the build fails on any of the Builders in the upstream set, the downstream builds will not fire. Note that, for SourceStamps generated by a Dependent scheduler, the revision is None, meaning HEAD. If any changes are committed between the time the upstream scheduler begins its build and the time the dependent scheduler begins its build, then those changes will be included in the downstream build. See the Triggerable scheduler for a more flexible dependency mechanism that can avoid this problem.

The keyword arguments to this scheduler are:



See Configuring Schedulers.
The upstream scheduler to watch. Note that this is an instance, not the name of the scheduler.


from buildbot.plugins import schedulers
tests = schedulers.SingleBranchScheduler(name="just-tests",
package = schedulers.Dependent(name="build-package",
                               upstream=tests, # <- no quotes!
                               builderNames=["make-tarball", "make-deb",
c['schedulers'] = [tests, package]
Periodic Scheduler

This simple scheduler just triggers a build every N seconds.

The arguments to this scheduler are:





This option only has effect when using multiple codebases. When True, it uses the last seen revision for each codebase that does not have a change. When False, the default value, codebases without changes will use the revision from the codebases argument.
If this is true, then builds will not be scheduled at the designated time unless the specified branch has seen an important change since the previous build.
See Configuring Schedulers.
The time, in seconds, after which to start a build.


from buildbot.plugins import schedulers
nightly = schedulers.Periodic(name="daily",
c['schedulers'] = [nightly]

The scheduler in this example just runs the full solaris build once per day. Note that this scheduler only lets you control the time between builds, not the absolute time-of-day of each Build, so this could easily wind up an evening or every afternoon scheduler depending upon when it was first activated.

Nightly Scheduler

This is highly configurable periodic build scheduler, which triggers a build at particular times of day, week, month, or year. The configuration syntax is very similar to the well-known crontab format, in which you provide values for minute, hour, day, and month (some of which can be wildcards), and a build is triggered whenever the current time matches the given constraints. This can run a build every night, every morning, every weekend, alternate Thursdays, on your boss's birthday, etc.

Pass some subset of minute, hour, dayOfMonth, month, and dayOfWeek; each may be a single number or a list of valid values. The builds will be triggered whenever the current time matches these values. Wildcards are represented by a '*' string. All fields default to a wildcard except 'minute', so with no fields this defaults to a build every hour, on the hour. The full list of parameters is:









This option only has effect when using multiple codebases. When True, it uses the last seen revision for each codebase that does not have a change. When False, the default value, codebases without changes will use the revision from the codebases argument.
If this is true, then builds will not be scheduled at the designated time unless the change filter has accepted an important change since the previous build.
(deprecated; use change_filter and codebases) The branch to build when the time comes, and the branch to filter for if change_filter is not specified. Remember that a value of None here means the default branch, and will not match other branches!
The minute of the hour on which to start the build. This defaults to 0, meaning an hourly build.
The hour of the day on which to start the build, in 24-hour notation. This defaults to *, meaning every hour.
The day of the month to start a build. This defaults to *, meaning every day.
The month in which to start the build, with January = 1. This defaults to *, meaning every month.
The day of the week to start a build, with Monday = 0. This defaults to *, meaning every day of the week.

For example, the following master.cfg clause will cause a build to be started every night at 3:00am:

from buildbot.plugins import schedulers
                       builderNames=['builder1', 'builder2'],
                       hour=3, minute=0))

This scheduler will perform a build each Monday morning at 6:23am and again at 8:23am, but only if someone has committed code in the interim:

                       dayOfWeek=0, hour=[6,8], minute=23,

The following runs a build every two hours, using Python's range function:

        branch=None, # default branch
        hour=range(0, 24, 2)))

Finally, this example will run only on December 24th:

        branch=None, # default branch
        builderNames=['flying_circuits', 'radar'],
Try Schedulers

This scheduler allows developers to use the buildbot try command to trigger builds of code they have not yet committed. See try for complete details.

Two implementations are available: Try_Jobdir and Try_Userpass. The former monitors a job directory, specified by the jobdir parameter, while the latter listens for PB connections on a specific port, and authenticates against userport.

The buildmaster must have a scheduler instance in the config file's schedulers list to receive try requests. This lets the administrator control who may initiate these trial builds, which branches are eligible for trial builds, and which Builders should be used for them.

The scheduler has various means to accept build requests. All of them enforce more security than the usual buildmaster ports do. Any source code being built can be used to compromise the worker accounts, but in general that code must be checked out from the VC repository first, so only people with commit privileges can get control of the workers. The usual force-build control channels can waste worker time but do not allow arbitrary commands to be executed by people who don't have those commit privileges. However, the source code patch that is provided with the trial build does not have to go through the VC system first, so it is important to make sure these builds cannot be abused by a non-committer to acquire as much control over the workers as a committer has. Ideally, only developers who have commit access to the VC repository would be able to start trial builds, but unfortunately the buildmaster does not, in general, have access to VC system's user list.

As a result, the try scheduler requires a bit more configuration. There are currently two ways to set this up:

jobdir (ssh)

This approach creates a command queue directory, called the jobdir, in the buildmaster's working directory. The buildmaster admin sets the ownership and permissions of this directory to only grant write access to the desired set of developers, all of whom must have accounts on the machine. The buildbot try command creates a special file containing the source stamp information and drops it in the jobdir, just like a standard maildir. When the buildmaster notices the new file, it unpacks the information inside and starts the builds.

The config file entries used by 'buildbot try' either specify a local queuedir (for which write and mv are used) or a remote one (using scp and ssh).

The advantage of this scheme is that it is quite secure, the disadvantage is that it requires fiddling outside the buildmaster config (to set the permissions on the jobdir correctly). If the buildmaster machine happens to also house the VC repository, then it can be fairly easy to keep the VC userlist in sync with the trial-build userlist. If they are on different machines, this will be much more of a hassle. It may also involve granting developer accounts on a machine that would not otherwise require them.

To implement this, the worker invokes ssh -l username host buildbot tryserver ARGS, passing the patch contents over stdin. The arguments must include the inlet directory and the revision information.

user+password (PB)

In this approach, each developer gets a username/password pair, which are all listed in the buildmaster's configuration file. When the developer runs buildbot try, their machine connects to the buildmaster via PB and authenticates themselves using that username and password, then sends a PB command to start the trial build.

The advantage of this scheme is that the entire configuration is performed inside the buildmaster's config file. The disadvantages are that it is less secure (while the cred authentication system does not expose the password in plaintext over the wire, it does not offer most of the other security properties that SSH does). In addition, the buildmaster admin is responsible for maintaining the username/password list, adding and deleting entries as developers come and go.

For example, to set up the jobdir style of trial build, using a command queue directory of MASTERDIR/jobdir (and assuming that all your project developers were members of the developers unix group), you would first set up that directory:

mkdir -p MASTERDIR/jobdir MASTERDIR/jobdir/new MASTERDIR/jobdir/cur MASTERDIR/jobdir/tmp
chgrp developers MASTERDIR/jobdir MASTERDIR/jobdir/*
chmod g+rwx,o-rwx MASTERDIR/jobdir MASTERDIR/jobdir/*

and then use the following scheduler in the buildmaster's config file:

from buildbot.plugins import schedulers
s = schedulers.Try_Jobdir(name="try1",
                          builderNames=["full-linux", "full-netbsd",
c['schedulers'] = [s]

Note that you must create the jobdir before telling the buildmaster to use this configuration, otherwise you will get an error. Also remember that the buildmaster must be able to read and write to the jobdir as well. Be sure to watch the twistd.log file (Logfiles) as you start using the jobdir, to make sure the buildmaster is happy with it.


Patches in the jobdir are encoded using netstrings, which place an arbitrary upper limit on patch size of 99999 bytes. If your submitted try jobs are rejected with BadJobfile, try increasing this limit with a snippet like this in your master.cfg:

from twisted.protocols.basic import NetstringReceiver
NetstringReceiver.MAX_LENGTH = 1000000

To use the username/password form of authentication, create a Try_Userpass instance instead. It takes the same builderNames argument as the Try_Jobdir form, but accepts an additional port argument (to specify the TCP port to listen on) and a userpass list of username/password pairs to accept. Remember to use good passwords for this: the security of the worker accounts depends upon it:

from buildbot.plugins import schedulers
s = schedulers.Try_Userpass(name="try2",
                            builderNames=["full-linux", "full-netbsd",
                            userpass=[("alice","pw1"), ("bob", "pw2")])
c['schedulers'] = [s]

Like most places in the buildbot, the port argument takes a strports specification. See twisted.application.strports for details.

Triggerable Scheduler

The Triggerable scheduler waits to be triggered by a Trigger step (see Triggering Schedulers) in another build. That step can optionally wait for the scheduler's builds to complete. This provides two advantages over Dependent schedulers. First, the same scheduler can be triggered from multiple builds. Second, the ability to wait for Triggerable's builds to complete provides a form of "subroutine call", where one or more builds can "call" a scheduler to perform some work for them, perhaps on other workers. The Triggerable scheduler supports multiple codebases. The scheduler filters out all codebases from Trigger steps that are not configured in the scheduler.

The parameters are just the basics:





See Configuring Schedulers.

This class is only useful in conjunction with the Trigger step. Here is a fully-worked example:

from buildbot.plugins import schedulers, util, steps

checkin = schedulers.SingleBranchScheduler(name="checkin",
nightly = schedulers.Nightly(name='nightly',
                             hour=3, minute=0)

mktarball = schedulers.Triggerable(name="mktarball", builderNames=["mktarball"])
build = schedulers.Triggerable(name="build-all-platforms",
test = schedulers.Triggerable(name="distributed-test",
package = schedulers.Triggerable(name="package-all-platforms",
c['schedulers'] = [mktarball, checkin, nightly, build, test, package]

# on checkin, make a tarball, build it, and test it
checkin_factory = util.BuildFactory()

# and every night, make a tarball, build it, and package it
nightly_factory = util.BuildFactory()
NightlyTriggerable Scheduler
class buildbot.schedulers.timed.NightlyTriggerable

The NightlyTriggerable scheduler is a mix of the Nightly and Triggerable schedulers. This scheduler triggers builds at a particular time of day, week, or year, exactly as the Nightly scheduler. However, the source stamp set that is used that provided by the last Trigger step that targeted this scheduler.

The parameters are just the basics:




See Configuring Schedulers.





See Nightly.

This class is only useful in conjunction with the Trigger step. Note that waitForFinish is ignored by Trigger steps targeting this scheduler.

Here is a fully-worked example:

from buildbot.plugins import schedulers, util, steps

checkin = schedulers.SingleBranchScheduler(name="checkin",
nightly = schedulers.NightlyTriggerable(name='nightly',
                                        hour=3, minute=0)
c['schedulers'] = [checkin, nightly]

# on checkin, run tests
checkin_factory = util.BuildFactory([

# and every night, package the latest successful build
nightly_factory = util.BuildFactory([
    steps.ShellCommand(command=['make', 'package'])
ForceScheduler Scheduler

The ForceScheduler scheduler is the way you can configure a force build form in the web UI.

In the /#/builders/:builderid web page, you will see, on the top right of the page, one button for each ForceScheduler scheduler that was configured for this builder. If you click on that button, a dialog will let you choose various parameters for requesting a new build.

The Buildbot framework allows you to customize exactly how the build form looks, which builders have a force build form (it might not make sense to force build every builder), and who is allowed to force builds on which builders.

How you do so is by configuring a ForceScheduler, and add it into the list schedulers.

The scheduler takes the following parameters:


Name of the scheduler (should be an Identifier).


List of builders where the force button should appear. See Configuring Schedulers.


A parameter allowing the user to specify the reason for the build. The default value is a string parameter with a default value "force build".


A string that will be used to create the build reason for the forced build. This string can contain the placeholders %(owner)s and %(reason)s, which represents the value typed into the reason field.


A parameter specifying the username associated with the build (aka owner). The default value is a username parameter.


A list of strings or CodebaseParameter specifying the codebases that should be presented. The default is a single codebase with no name (i.e. codebases=['']).


A list of parameters, one for each property. These can be arbitrary parameters, where the parameter's name is taken as the property name, or AnyPropertyParameter, which allows the web user to specify the property name. The default value is an empty list.


The name of the "submit" button on the resulting force-build form. This defaults to the name of scheduler.

An example may be better than long explanation. What you need in your config file is something like:

from buildbot.plugins import schedulers, util

sch = schedulers.ForceScheduler(
    label="My nice Force form",

            name="Main repository",
            # will generate a combo box
                choices=["master", "hest"],

            # will generate nothing in the form, but revision, repository,
            # and project are needed by buildbot scheduling system so we
            # need to pass a value ("")
            revision=util.FixedParameter(name="revision", default=""),
            repository=util.FixedParameter(name="repository", default=""),
            project=util.FixedParameter(name="project", default=""),

    # will generate a text input
                                required=True, size=80),

    # in case you dont require authentication this will display
    # input for user to type his name
    username=util.UserNameParameter(label="your name:",
    # A completely customized property list.  The name of the
    # property is the name of the parameter
        util.NestedParameter(name="options", label="Build Options", layout="vertical", fields=[
                                 label="optionally give a public Git pull url:",
                                 default="", size=80),
                                  label="force a make clean",

This will result in the following UI:

Force Form Result

The force scheduler uses the web interface's authorization framework to determine which user has the right to force which build. Here is an example of code on how you can define which user has which right:

user_mapping = {
    re.compile("project1-builder"): ["project1-maintainer", "john"] ,
    re.compile("project2-builder"): ["project2-maintainer", "jack"],
    re.compile(".*"): ["root"]
def force_auth(user,  status):
    global user_mapping
    for r,users in user_mapping.items():
        if r.match(
            if user in users:
                    return True
    return False

# use authz_cfg in your WebStatus setup
    forceBuild = force_auth,
ForceScheduler Parameters

Most of the arguments to ForceScheduler are "parameters". Several classes of parameters are available, each describing a different kind of input from a force-build form.

All parameter types have a few common arguments:

name (required)

The name of the parameter. For properties, this will correspond to the name of the property that your parameter will set. The name is also used internally as the identifier for in the HTML form.

label (optional; default is same as name)

The label of the parameter. This is what is displayed to the user.

tablabel (optional; default is same as label)

The label of the tab if this parameter is included into a tab layout NestedParameter. This is what is displayed to the user.

default (optional; default: "")

The default value for the parameter, that is used if there is no user input.

required (optional; default: False)

If this is true, then an error will be shown to user if there is no input in this field

The parameter types are:

NestedParameter(name="options", label="Build options" layout="vertical", fields=[...]),

This parameter type is a special parameter which contains other parameters. This can be used to group a set of parameters together, and define the layout of your form. You can recursively include NestedParameter into NestedParameter, to build very complex UI.

It adds the following arguments:

layout (optional, default: "vertical")

The layout defines how the fields are placed in the form.

The layouts implemented in the standard web application are:

  • simple: fields are displayed one by one without alignment.

    They take the horizontal space that they need.

  • vertical: all fields are displayed vertically, aligned in columns (as per the column attribute of the NestedParameter)

  • tabs: Each field gets its own tab.

    This can be used to declare complex build forms which won't fit into one screen. The children fields are usually other NestedParameters with vertical layout.

columns (optional, accepted values are 1,2,3,4)

The number of columns to use for a vertical layout. If omitted, it is set to 1 unless there are more than 3 visible child fields in which case it is set to 2.
FixedParameter(name="branch", default="trunk"),

This parameter type will not be shown on the web form, and always generate a property with its default value.

    label="optionally give a public Git pull url:",
    default="", size=80)

This parameter type will show a single-line text-entry box, and allow the user to enter an arbitrary string. It adds the following arguments:

regex (optional)

A string that will be compiled as a regex, and used to validate the input of this parameter.

size (optional; default: 10)

The width of the input field (in characters).
    label="comments to be displayed to the user of the built binary",
    default="This is a development build", cols=60, rows=5)

This parameter type is similar to StringParameter, except that it is represented in the HTML form as a textarea, allowing multi-line input. It adds the StringParameter arguments, this type allows:

cols (optional; default: 80)

The number of columns the textarea will have.

rows (optional; default: 20)

The number of rows the textarea will have

This class could be subclassed in order to have more customization e.g.

  • developer could send a list of Git branches to pull from
  • developer could send a list of gerrit changes to cherry-pick,
  • developer could send a shell script to amend the build.

Beware of security issues anyway.

    label="debug level (1-10)", default=2)

This parameter type accepts an integer value using a text-entry box.

    label="force a make clean", default=False)

This type represents a boolean value. It will be presented as a checkbox.

UserNameParameter(label="your name:", size=80)

This parameter type accepts a username. If authentication is active, it will use the authenticated user instead of displaying a text-entry box.

size (optional; default: 10)
The width of the input field (in characters).
need_email (optional; default True)
If true, require a full email address rather than arbitrary text.
    choices=["main","devel"], default="main")

This parameter type lets the user choose between several choices (e.g the list of branches you are supporting, or the test campaign to run). If multiple is false, then its result is a string - one of the choices. If multiple is true, then the result is a list of strings from the choices.

Note that for some use cases, the choices need to be generated dynamically. This can be done via subclassing and overriding the 'getChoices' member function. An example of this is provided by the source for the InheritBuildParameter class.

Its arguments, in addition to the common options, are:


The list of available choices.

strict (optional; default: True)

If true, verify that the user's input is from the list. Note that this only affects the validation of the form request; even if this argument is False, there is no HTML form component available to enter an arbitrary value.


If true, then the user may select multiple choices.


                      label="smoke test campaign to run",
                      choices=["test_builder1", "test_builder2",

# .. and later base the schedulers to trigger off this property:

# triggers the tests depending on the property forced_test
builder1.factory.addStep(Trigger(name="Trigger tests",

This is a parameter group to specify a sourcestamp for a given codebase.


The name of the codebase.

branch (optional; default: StringParameter)

A parameter specifying the branch to build. The default value is a string parameter.

revision (optional; default: StringParameter)

A parameter specifying the revision to build. The default value is a string parameter.

repository (optional; default: StringParameter)

A parameter specifying the repository for the build. The default value is a string parameter.

project (optional; default: StringParameter)

A parameter specifying the project for the build. The default value is a string parameter.


InheritBuildParameter is not yet ported to data API, and cannot be used with buildbot nine yet(bug #3521).

This is a special parameter for inheriting force build properties from another build. The user is presented with a list of compatible builds from which to choose, and all forced-build parameters from the selected build are copied into the new build. The new parameter is:


A function to find compatible builds in the build history. This function is given the master Status instance as first argument, and the current builder name as second argument, or None when forcing all builds.


def get_compatible_builds(status, builder):
    if builder is None: # this is the case for force_build_all
        return ["cannot generate build list here"]
    # find all successful builds in builder1 and builder2
    builds = []
    for builder in ["builder1","builder2"]:
        builder_status = status.getBuilder(builder)
        for num in xrange(1,40): # 40 last builds
            b = builder_status.getBuild(-num)
            if not b:
            if b.getResults() == FAILURE:
    return builds

# ...

sched = Scheduler(...,
            label="promote a build for merge",
            required = True),


WorkerChoiceParameter is not yet ported to data API, and cannot be used with buildbot nine yet(bug #3521).

This parameter allows a scheduler to require that a build is assigned to the chosen worker. The choice is assigned to the workername property for the build. The enforceChosenWorker functor must be assigned to the canStartBuild parameter for the Builder.


from buildbot.plugins import util

# schedulers:
    # ...

# builders:
    # ...

This parameter type can only be used in properties, and allows the user to specify both the property name and value in the web form.

This Parameter is here to reimplement old Buildbot behavior, and should be avoided. Stricter parameter name and type should be preferred.

2.4.5. Workers

The workers configuration key specifies a list of known workers. In the common case, each worker is defined by an instance of the Worker class. It represents a standard, manually started machine that will try to connect to the buildbot master as a worker. Buildbot also supports "on-demand", or latent, workers, which allow buildbot to dynamically start and stop worker instances. Defining Workers

A Worker instance is created with a workername and a workerpassword. These are the same two values that need to be provided to the worker administrator when they create the worker.

The workername must be unique, of course. The password exists to prevent evildoers from interfering with the buildbot by inserting their own (broken) workers into the system and thus displacing the real ones.

Workers with an unrecognized workername or a non-matching password will be rejected when they attempt to connect, and a message describing the problem will be written to the log file (see Logfiles).

A configuration for two workers would look like:

from buildbot.plugins import worker
c['workers'] = [
    worker.Worker('bot-solaris', 'solarispasswd'),
    worker.Worker('bot-bsd', 'bsdpasswd'),
] Worker Options

Worker objects can also be created with an optional properties argument, a dictionary specifying properties that will be available to any builds performed on this worker. For example:

c['workers'] = [
    worker.Worker('bot-solaris', 'solarispasswd',
                  properties={ 'os':'solaris' }),

The Worker constructor can also take an optional max_builds parameter to limit the number of builds that it will execute simultaneously:

c['workers'] = [
    worker.Worker("bot-linux", "linuxpassword", max_builds=2)
Master-Worker TCP Keepalive

By default, the buildmaster sends a simple, non-blocking message to each worker every hour. These keepalives ensure that traffic is flowing over the underlying TCP connection, allowing the system's network stack to detect any problems before a build is started.

The interval can be modified by specifying the interval in seconds using the keepalive_interval parameter of Worker:

c['workers'] = [
    worker.Worker('bot-linux', 'linuxpasswd',

The interval can be set to None to disable this functionality altogether.

When Workers Go Missing

Sometimes, the workers go away. One very common reason for this is when the worker process is started once (manually) and left running, but then later the machine reboots and the process is not automatically restarted.

If you'd like to have the administrator of the worker (or other people) be notified by email when the worker has been missing for too long, just add the notify_on_missing= argument to the Worker definition. This value can be a single email address, or a list of addresses:

c['workers'] = [
    worker.Worker('bot-solaris', 'solarispasswd',

By default, this will send email when the worker has been disconnected for more than one hour. Only one email per connection-loss event will be sent. To change the timeout, use missing_timeout= and give it a number of seconds (the default is 3600).

You can have the buildmaster send email to multiple recipients: just provide a list of addresses instead of a single one:

c['workers'] = [
    worker.Worker('bot-solaris', 'solarispasswd',
                  missing_timeout=300)  # notify after 5 minutes

The email sent this way will use a MailNotifier (see MailNotifier) status target, if one is configured. This provides a way for you to control the from address of the email, as well as the relayhost (aka smarthost) to use as an SMTP server. If no MailNotifier is configured on this buildmaster, the worker-missing emails will be sent using a default configuration.

Note that if you want to have a MailNotifier for worker-missing emails but not for regular build emails, just create one with builders=[], as follows:

from buildbot.plugins import status, worker
m = status.MailNotifier(fromaddr="buildbot@localhost", builders=[],

c['workers'] = [
        worker.Worker('bot-solaris', 'solarispasswd',
] Local Workers

For smaller setups, you may want to just run the workers on the same machine as the master. To simplify the maintainance, you may even want to run them in the same process.

This is what LocalWorker is for. Instead of configuring a worker.Worker, you have to configure a worker.LocalWorker. As the worker is running on the same process, password is not necessary. You can run as many local workers as long as your machine CPU and memory is allowing.

A configuration for two workers would look like:

from buildbot.plugins import worker
c['workers'] = [

In order to use local workers you need to have buildbot-worker package installed. Latent Workers

The standard buildbot model has workers started manually. The previous section described how to configure the master for this approach.

Another approach is to let the buildbot master start workers when builds are ready, on-demand. Thanks to services such as Amazon Web Services' Elastic Compute Cloud ("AWS EC2"), this is relatively easy to set up, and can be very useful for some situations.

The workers that are started on-demand are called "latent" workers. As of this writing, buildbot ships with an abstract base class for building latent workers, and a concrete implementation for AWS EC2 and for libvirt.

Common Options

The following options are available for all latent workers.

This option allows you to specify how long a latent worker should wait after a build for another build before it shuts down. It defaults to 10 minutes. If this is set to 0 then the worker will be shut down immediately. If it is less than 0 it will never automatically shutdown.
Supported Latent Workers

As of time of writing, Buildbot supports the following latent workers:

Amazon Web Services Elastic Compute Cloud ("AWS EC2")
class buildbot.worker.ec2.EC2LatentWorker

EC2 is a web service that allows you to start virtual machines in an Amazon data center. Please see their website for details, including costs. Using the AWS EC2 latent workers involves getting an EC2 account with AWS and setting up payment; customizing one or more EC2 machine images ("AMIs") on your desired operating system(s) and publishing them (privately if needed); and configuring the buildbot master to know how to start your customized images for "substantiating" your latent workers.

This document will guide you through setup of a AWS EC2 latent worker:

Get an AWS EC2 Account

To start off, to use the AWS EC2 latent worker, you need to get an AWS developer account and sign up for EC2. Although Amazon often changes this process, these instructions should help you get started:

  1. Go to and click to "Sign Up Now" for an AWS account.
  2. Once you are logged into your account, you need to sign up for EC2. Instructions for how to do this have changed over time because Amazon changes their website, so the best advice is to hunt for it. After signing up for EC2, it may say it wants you to upload an x.509 cert. You will need this to create images (see below) but it is not technically necessary for the buildbot master configuration.
  3. You must enter a valid credit card before you will be able to use EC2. Do that under 'Payment Method'.
  4. Make sure you're signed up for EC2 by going to Your Account ‣ Account Activity and verifying EC2 is listed.
Create an AMI

Now you need to create an AMI and configure the master. You may need to run through this cycle a few times to get it working, but these instructions should get you started.

Creating an AMI is out of the scope of this document. The EC2 Getting Started Guide is a good resource for this task. Here are a few additional hints.

  • When an instance of the image starts, it needs to automatically start a buildbot worker that connects to your master (to create a buildbot worker, Creating a worker; to make a daemon, Launching the daemons).
  • You may want to make an instance of the buildbot worker, configure it as a standard worker in the master (i.e., not as a latent worker), and test and debug it that way before you turn it into an AMI and convert to a latent worker in the master.
Configure the Master with an EC2LatentWorker

Now let's assume you have an AMI that should work with the EC2LatentWorker. It's now time to set up your buildbot master configuration.

You will need some information from your AWS account: the Access Key Id and the Secret Access Key. If you've built the AMI yourself, you probably already are familiar with these values. If you have not, and someone has given you access to an AMI, these hints may help you find the necessary values:

  • While logged into your AWS account, find the "Access Identifiers" link (either on the left, or via Your Account ‣ Access Identifiers.
  • On the page, you'll see alphanumeric values for "Your Access Key Id:" and "Your Secret Access Key:". Make a note of these. Later on, we'll call the first one your identifier and the second one your secret_identifier.

When creating an EC2LatentWorker in the buildbot master configuration, the first three arguments are required. The name and password are the first two arguments, and work the same as with normal workers. The next argument specifies the type of the EC2 virtual machine (available options as of this writing include m1.small, m1.large, m1.xlarge, c1.medium, and c1.xlarge; see the EC2 documentation for descriptions of these machines).

Here is the simplest example of configuring an EC2 latent worker. It specifies all necessary remaining values explicitly in the instantiation.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

The ami argument specifies the AMI that the master should start. The identifier argument specifies the AWS Access Key Id, and the secret_identifier specifies the AWS Secret Access Key. Both the AMI and the account information can be specified in alternate ways.


Whoever has your identifier and secret_identifier values can request AWS work charged to your account, so these values need to be carefully protected. Another way to specify these access keys is to put them in a separate file. Buildbot supports the standard AWS credentials file. You can then make the access privileges stricter for this separate file, and potentially let more people read your main configuration file. If your master is running in EC2, you can also use IAM roles for EC2 to delegate permissions.

keypair_name and security_name allow you to specify different names for these AWS EC2 values.

You can make an .aws directory in the home folder of the user running the buildbot master. In that directory, create a file called credentials. The format of the file should be as follows, replacing identifier and secret_identifier with the credentials obtained before.

aws_access_key_id = identifier
aws_secret_access_key = secret_identifier

If you are using IAM roles, no config file is required. Then you can instantiate the worker as follows.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

Previous examples used a particular AMI. If the Buildbot master will be deployed in a process-controlled environment, it may be convenient to specify the AMI more flexibly. Rather than specifying an individual AMI, specify one or two AMI filters.

In all cases, the AMI that sorts last by its location (the S3 bucket and manifest name) will be preferred.

One available filter is to specify the acceptable AMI owners, by AWS account number (the 12 digit number, usually rendered in AWS with hyphens like "1234-5678-9012", should be entered as in integer).

from buildbot.plugins import worker
bot1 = worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

The other available filter is to provide a regular expression string that will be matched against each AMI's location (the S3 bucket and manifest name).

from buildbot.plugins import worker
bot1 = worker.EC2LatentWorker(
        'bot1', 'sekrit', 'm1.large',

The regular expression can specify a group, which will be preferred for the sorting. Only the first group is used; subsequent groups are ignored.

from buildbot.plugins import worker
bot1 = worker.EC2LatentWorker(
    'bot1', 'sekrit', 'm1.large',

If the group can be cast to an integer, it will be. This allows 10 to sort after 1, for instance.

from buildbot.plugins import worker
bot1 = worker.EC2LatentWorker(
        'bot1', 'sekrit', 'm1.large',

In addition to using the password as a handshake between the master and the worker, you may want to use a firewall to assert that only machines from a specific IP can connect as workers. This is possible with AWS EC2 by using the Elastic IP feature. To configure, generate a Elastic IP in AWS, and then specify it in your configuration using the elastic_ip argument.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

One other way to configure a worker is by settings AWS tags. They can for example be used to have a more restrictive security IAM policy. To get Buildbot to tag the latent worker specify the tag keys and values in your configuration using the tags argument.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',
                           tags={'SomeTag': 'foo'})

If the worker needs access to additional AWS resources, you can also enable your workers to access them via an EC2 instance profile. To use this capability, you must first create an instance profile separately in AWS. Then specify its name on EC2LatentWorker via instance_profile_name.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

You may also supply your own boto3.Session object to allow for more flexible session options (ex. cross-account) To use this capability, you must first create a boto3.Session object. Then provide it to EC2LatentWorker via session argument.

import boto3
from buildbot.plugins import worker

session = boto3.session.Session()
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',

The EC2LatentWorker supports all other configuration from the standard Worker. The missing_timeout and notify_on_missing specify how long to wait for an EC2 instance to attach before considering the attempt to have failed, and email addresses to alert, respectively. missing_timeout defaults to 20 minutes.


If you want to attach existing volumes to an ec2 latent worker, use the volumes attribute. This mechanism can be valuable if you want to maintain state on a conceptual worker across multiple start/terminate sequences. volumes expects a list of (volume_id, mount_point) tuples to attempt attaching when your instance has been created.

If you want to attach new ephemeral volumes, use the the block_device_map attribute. This follows the AWS API syntax, essentially acting as a passthrough. The only distinction is that the volumes default to deleting on termination to avoid leaking volume resources when workers are terminated. See boto documentation for further details.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',
                           block_device_map= [
                             "DeviceName": "/dev/xvdb",
                             "Ebs" : {
                                "VolumeType": "io1",
                                "Iops": 1000,
                                "VolumeSize": 100
VPC Support

If you are managing workers within a VPC, your worker configuration must be modified from above. You must specify the id of the subnet where you want your worker placed. You must also specify security groups created within your VPC as opposed to classic EC2 security groups. This can be done by passing the ids of the vpc security groups. Note, when using a VPC, you can not specify classic EC2 security groups (as specified by security_name).

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',
Spot instances

If you would prefer to use spot instances for running your builds, you can accomplish that by passing in a True value to the spot_instance parameter to the EC2LatentWorker constructor. Additionally, you may want to specify max_spot_price and price_multiplier in order to limit your builds' budget consumption.

from buildbot.plugins import worker
c['workers'] = [
    worker.EC2LatentWorker('bot1', 'sekrit', 'm1.large',
                           'ami-12345', region='us-west-2',
                           placement='b', spot_instance=True,

This example would attempt to create a m1.large spot instance in the us-west-2b region costing no more than $0.09/hour. The spot prices for 'Linux/UNIX' spot instances in that region over the last 24 hours will be averaged and multiplied by the price_multiplier parameter, then a spot request will be sent to Amazon with the above details. If the multiple exceeds the max_spot_price, the bid price will be the max_spot_price.

Either max_spot_price or price_multiplier, but not both, may be None. If price_multiplier is None, then no historical price information is retrieved; the bid price is simply the specified max_spot_price. If the max_spot_price is None, then the multiple of the historical average spot prices is used as the bid price with no limit.

class buildbot.worker.libvirt.LibVirtWorker

libvirt is a virtualization API for interacting with the virtualization capabilities of recent versions of Linux and other OSes. It is LGPL and comes with a stable C API, and Python bindings.

This means we know have an API which when tied to buildbot allows us to have workers that run under Xen, QEMU, KVM, LXC, OpenVZ, User Mode Linux, VirtualBox and VMWare.

The libvirt code in Buildbot was developed against libvirt 0.7.5 on Ubuntu Lucid. It is used with KVM to test Python code on Karmic VM's, but obviously isn't limited to that. Each build is run on a new VM, images are temporary and thrown away after each build.

This document will guide you through setup of a libvirt latent worker:

Setting up libvirt

We won't show you how to set up libvirt as it is quite different on each platform, but there are a few things you should keep in mind.

  • If you are running on Ubuntu, your master should run Lucid. Libvirt and apparmor are buggy on Karmic.
  • If you are using the system libvirt, your buildbot master user will need to be in the libvirtd group.
  • If you are using KVM, your buildbot master user will need to be in the KVM group.
  • You need to think carefully about your virtual network first. Will NAT be enough? What IP will my VM's need to connect to for connecting to the master?
Configuring your base image

You need to create a base image for your builds that has everything needed to build your software. You need to configure the base image with a buildbot worker that is configured to connect to the master on boot.

Because this image may need updating a lot, we strongly suggest scripting its creation.

If you want to have multiple workers using the same base image it can be annoying to duplicate the image just to change the buildbot credentials. One option is to use libvirt's DHCP server to allocate an identity to the worker: DHCP sets a hostname, and the worker takes its identity from that.

Doing all this is really beyond the scope of the manual, but there is a vmbuilder script and a network.xml file to create such a DHCP server in contrib/ (Contrib Scripts) that should get you started:

sudo apt-get install ubuntu-vm-builder
sudo contrib/libvirt/vmbuilder

Should create an ubuntu/ folder with a suitable image in it.

virsh net-define contrib/libvirt/network.xml
virsh net-start buildbot-network

Should set up a KVM compatible libvirt network for your buildbot VM's to run on.

Configuring your Master

If you want to add a simple on demand VM to your setup, you only need the following. We set the username to minion1, the password to sekrit. The base image is called base_image and a copy of it will be made for the duration of the VM's life. That copy will be thrown away every time a build is complete.

from buildbot.plugins import worker, util
c['workers'] = [
    worker.LibVirtWorker('minion1', 'sekrit',

You can use virt-manager to define minion1 with the correct hardware. If you don't, buildbot won't be able to find a VM to start.

LibVirtWorker accepts the following arguments:

Both a buildbot username and the name of the virtual machine.
A password for the buildbot to login to the master with.
Connection instance wrapping connection to libvirt.
The path to a libvirt disk image, normally in qcow2 format when using KVM.
If given a base image, buildbot will clone it every time it starts a VM. This means you always have a clean environment to do your build in.
If a VM isn't predefined in virt-manager, then you can instead provide XML like that used with virsh define. The VM will be created automatically when needed, and destroyed when not needed any longer.
class buildbot.worker.openstack.OpenStackLatentWorker

OpenStack is a series of interconnected components that facilitates managing compute, storage, and network resources in a data center. It is available under the Apache License and has a REST interface along with a Python client.

This document will guide you through setup of an OpenStack latent worker:

Install dependencies

OpenStackLatentWorker requires python-novaclient to work, you can install it with pip install python-novaclient.

Get an Account in an OpenStack cloud

Setting up OpenStack is outside the domain of this document. There are four account details necessary for the Buildbot master to interact with your OpenStack cloud: username, password, a tenant name, and the auth URL to use.

Create an Image

OpenStack supports a large number of image formats. OpenStack maintains a short list of prebuilt images; if the desired image is not listed, The OpenStack Compute Administration Manual is a good resource for creating new images. You need to configure the image with a buildbot worker to connect to the master on boot.

Configure the Master with an OpenStackLatentWorker

With the configured image in hand, it is time to configure the buildbot master to create OpenStack instances of it. You will need the aforementioned account details. These are the same details set in either environment variables or passed as options to an OpenStack client.

OpenStackLatentWorker accepts the following arguments:

The worker name.
A password for the worker to login to the master with.
The flavor ID to use for the instance.
A string containing the image UUID to use for the instance. A callable may instead be passed. It will be passed the list of available images and must return the image to use.




The OpenStack authentication needed to create and delete instances. These are the same as the environment variables with uppercase names of the arguments.

A list of dictionaries. Each dictionary specifies a block device to set up during instance creation.

Supported keys

(required): The image, snapshot, or volume UUID.
(required): Size of the block device in GiB.
(optional): defaults to vda. The name of the device in the instance; e.g. vda or xda.
(optional): defaults to image. The origin of the block device. Valid values are image, snapshot, or volume.
(optional): defaults to volume. Destination of block device: volume or local.
(optional): defaults to True. Controls if the block device will be deleted when the instance terminates.
(optional): defaults to 0. Integer used for boot order.
A dictionary of string key-value pairs to pass to the instance. These will be available under the metadata key from the metadata service.
(optional) A dict that will be appended to the arguments when creating a VM. Buildbot uses the OpenStack Nova version 1.1 API by default (see client_version).
(optional) Nova client version to use. Defaults to 1.1 (deprecated). Use 2 or 2.minor for version 2 API.

Here is the simplest example of configuring an OpenStack latent worker.

from buildbot.plugins import worker
c['workers'] = [
    worker.OpenStackLatentWorker('bot2', 'sekrit',
                flavor=1, image='8ac9d4a4-5e03-48b0-acde-77a0345a9ab1',
                os_username='user', os_password='password',

The image argument also supports being given a callable. The callable will be passed the list of available images and must return the image to use. The invocation happens in a separate thread to prevent blocking the build master when interacting with OpenStack.

from buildbot.plugins import worker

def find_image(images):
    # Sort oldest to newest.
    cmp_fn = lambda x,y: cmp(x.created, y.created)
    candidate_images = sorted(images, cmp=cmp_fn)
    # Return the oldest candiate image.
    return candidate_images[0]

c['workers'] = [
    worker.OpenStackLatentWorker('bot2', 'sekrit',
                flavor=1, image=find_image,
                os_username='user', os_password='password',

The block_devices argument is minimally manipulated to provide some defaults and passed directly to novaclient. The simplest example is an image that is converted to a volume and the instance boots from that volume. When the instance is destroyed, the volume will be terminated as well.

from buildbot.plugins import worker
c['workers'] = [
    worker.OpenStackLatentWorker('bot2', 'sekrit',
                flavor=1, image='8ac9d4a4-5e03-48b0-acde-77a0345a9ab1',
                os_username='user', os_password='password',
                    {'uuid': '3f0b8868-67e7-4a5b-b685-2824709bd486',
                    'volume_size': 10}])

The nova_args can be used to specify additional arguments for the novaclient. For example network mappings, which is required if your OpenStack tenancy has more than one network, and default cannot be determined. Please refer to your OpenStack manual whether it wants net-id or net-name.

Other useful parameters are availability_zone, security_groups and config_drive. Refer to Python bindings to the OpenStack Nova API for more information. It is found on section Servers, method create.

from buildbot.plugins import worker
c['workers'] = [
    worker.OpenStackLatentWorker('bot2', 'sekrit',
                flavor=1, image='8ac9d4a4-5e03-48b0-acde-77a0345a9ab1',
                os_username='user', os_password='password',
                  'nics': [

OpenStackLatentWorker supports all other configuration from the standard Worker. The missing_timeout and notify_on_missing specify how long to wait for an OpenStack instance to attach before considering the attempt to have failed and email addresses to alert, respectively. missing_timeout defaults to 20 minutes.

Docker latent worker
class buildbot.worker.docker.DockerLatentWorker

Docker is an open-source project that automates the deployment of applications inside software containers. Using the Docker latent worker, an attempt is made at instantiating a fresh image upon each build, assuring consistency of the environment between builds. Each image will be discarded once the worker finished processing the build queue (i.e. becomes idle). See build_wait_timeout to change this behavior.

This document will guide you through the setup of such workers.

Docker Installation

An easy way to try Docker is through installation of dedicated Virtual machines. Two of them stands out:

Beside, it is always possible to install Docker next to the buildmaster. Beware that in this case, overall performance will depend on how many builds the computer where you have your buildmaster can handle as everything will happen on the same one.


It is not necessary to install Docker in the same environment as your master as we will make use to the Docker API through docker-py. More in master setup.


CoreOS is targeted at building infrastructure and distributed systems. In order to get the latent worker working with CoreOS, it is necessary to expose the docker socket outside of the Virtual Machine. If you installed it via Vagrant, it is also necessary to uncomment the following line in your config.rb file:


The following command should allow you to confirm that your Docker socket is now available via the network:

docker -H tcp:// ps

boot2docker is one of the fastest ways to boot to Docker. As it is meant to be used from outside of the Virtual Machine, the socket is already exposed. Please follow the installation instructions on how to find the address of your socket.

Image Creation

Our build master will need the name of an image to perform its builds. Each time a new build will be requested, the same base image will be used again and again, actually discarding the result of the previous build. If you need some persistant storage between builds, you can use Volumes.

Each Docker image has a single purpose. Our worker image will be running a buildbot worker.

Docker uses Dockerfiles to describe the steps necessary to build an image. The following example will build a minimal worker. Don't forget to add your dependencies in there to get a succesfull build !

FROM debian:stable
RUN apt-get update && apt-get install -y \
   python-dev \
RUN pip install buildbot-worker
RUN groupadd -r buildbot && useradd -r -g buildbot buildbot
RUN mkdir /worker && chown buildbot:buildbot /worker
# Install your build-dependencies here ...
USER buildbot
WORKDIR /worker
RUN buildbot-worker create-worker . <master-hostname> <workername> <workerpassword>
ENTRYPOINT ["/usr/local/bin/buildbot-worker"]
CMD ["start", "--nodaemon"]

On line 11, the hostname for your master instance, as well as the worker name and password is setup. Don't forget to replace those values with some valid ones for your project.

It is a good practice to set the ENTRYPOINT to the worker executable, and the CMD to ["start", "--nodaemon"]. This way, no parameter will be required when starting the image.

When your Dockerfile is ready, you can build your first image using the following command (replace myworkername with a relevant name for your case):

docker build -t myworkername - < Dockerfile
Reuse same image for different workers

Previous simple example hardcodes the worker name into the dockerfile, which will not work if you want to share your docker image between workers.

You can find in buildbot source code in master/contrib/docker two example configurations:

the base worker configuration, including a custom buildbot.tac, which takes environment variables into account for setting the correct worker name, and connect to the correct master.
a worker with Python and node installed, which demonstrate how to reuse the base worker to create variations of build environments.

The master setups several environment variables before starting the workers:

The address of the master the worker shall connect to
The port of the master's worker 'pb' protocol.
The name the worker should use to connect to master
The password the worker should use to connect to master
Master Setup

We will rely on docker-py to connect our master with docker. Now is the time to install it in your master environment.

Before adding the worker to your master configuration, it is possible to validate the previous steps by starting the newly created image interactively. To do this, enter the following lines in a Python prompt where docker-py is installed:

>>> import docker
>>> docker_socket = 'tcp://localhost:2375'
>>> client = docker.client.Client(base_url=docker_socket)
>>> worker_image = 'my_project_worker'
>>> container = client.create_container(worker_image)
>>> client.start(container['Id'])
>>> # Optionally examine the logs of the master
>>> client.stop(container['Id'])
>>> client.wait(container['Id'])

It is now time to add the new worker to the master configuration under workers.

The following example will add a Docker latent worker for docker running at the following adress: tcp://localhost:2375, the worker name will be docker, its password: password, and the base image name will be my_project_worker:

from buildbot.plugins import worker
c['workers'] = [
    worker.DockerLatentWorker('docker', 'password',

In addition to the arguments available for any Latent Workers, DockerLatentWorker will accept the following extra ones:

(mandatory) This is the adress the master will use to connect with a running Docker instance.
(optional if dockerfile is given) This is the name of the image that will be started by the build master. It should start a worker. This option can be a renderable, like Interpolate, so that it generates from the build request properties.
(optional) This will override the command setup during image creation.
(optional) See Setting up Volumes

(optional if image is given) This is the content of the Dockerfile that will be used to build the specified image if the image is not found by Docker. It should be a multiline string.


In case image and dockerfile are given, no attempt is made to compare the image with the content of the Dockerfile parameter if the image is found.

(optional, default to the highest version known by docker-py) This will indicates wich API version must be used to communicate with Docker.
(optional) This allow to use TLS when connecting with the Docker socket. This should be a docker.tls.TLSConfig object. See docker-py's own documentation for more details on how to initialise this object.
(optional, defaults to false) This transfers docker container's log inside master logs during worker startup (before connection). This can be useful to debug worker startup. e.g network issues, etc.
(optional, defaults to socket.getfqdn()) Address of the master the worker should connect to. Use if you master machine does not have proper fqdn. This value is passed to the docker image via environment variable BUILDMASTER
(optional) Extra host configuration parameters passed as a dictionary used to create HostConfig object. See docker-py's HostConfig documentation for all the supported options.
Setting up Volumes

The volume parameter allows to share directory between containers, or between a container and the host system. Refer to Docker documentation for more information about Volumes.

The format of that variable has to be an array of string. Each string specify a volume in the following format: volumename:bindname. The volume name has to be appended with :ro if the volume should be mounted read-only.


This is the same format as when specifying volumes on the command line for docker's own -v option.

Dangers with Latent Workers

Any latent worker that interacts with a for-fee service, such as the EC2LatentWorker, brings significant risks. As already identified, the configuration will need access to account information that, if obtained by a criminal, can be used to charge services to your account. Also, bugs in the buildbot software may lead to unnecessary charges. In particular, if the master neglects to shut down an instance for some reason, a virtual machine may be running unnecessarily, charging against your account. Manual and/or automatic (e.g. nagios with a plugin using a library like boto) double-checking may be appropriate.

A comparatively trivial note is that currently if two instances try to attach to the same latent worker, it is likely that the system will become confused. This should not occur, unless, for instance, you configure a normal worker to connect with the authentication of a latent buildbot. If this situation does occurs, stop all attached instances and restart the master.

2.4.6. Builder Configuration

The builders configuration key is a list of objects giving configuration for the Builders. For more information on the function of Builders in Buildbot, see the Concepts chapter. The class definition for the builder configuration is in buildbot.config. However there is a much simpler way to use it, so in the configuration file, its use looks like:

from buildbot.plugins import util
c['builders'] = [
    util.BuilderConfig(name='quick', workernames=['bot1', 'bot2'], factory=f_quick),
    util.BuilderConfig(name='thorough', workername='bot1', factory=f_thorough),

BuilderConfig takes the following keyword arguments:

This specifies the Builder's name, which is used in status reports.


These arguments specify the worker or workers that will be used by this Builder. All workers names must appear in the workers configuration parameter. Each worker can accommodate multiple builders. The workernames parameter can be a list of names, while workername can specify only one worker.
This is a buildbot.process.factory.BuildFactory instance which controls how the build is performed by defining the steps in the build. Full details appear in their own section, Build Factories.

Other optional keys may be set on each BuilderConfig:

Specifies the name of a subdirectory of the master's basedir in which everything related to this builder will be stored. This holds build status information. If not set, this parameter defaults to the builder name, with some characters escaped. Each builder must have a unique build directory.
Specifies the name of a subdirectory (under the worker's configured base directory) in which everything related to this builder will be placed on the worker. This is where checkouts, compiles, and tests are run. If not set, defaults to builddir. If a worker is connected to multiple builders that share the same workerbuilddir, make sure the worker is set to run one build at a time or ensure this is fine to run multiple builds from the same directory simultaneously.
If provided, this is a list of strings that identifies tags for the builder. Status clients can limit themselves to a subset of the available tags. A common use for this is to add new builders to your setup (for a new module, or for a new worker) that do not work correctly yet and allow you to integrate them with the active builders. You can tag these new builders with a test tag, make your main status clients ignore them, and have only private status clients pick them up. As soon as they work, you can move them over to the active tag.
If provided, this is a function that controls which worker will be assigned future jobs. The function is passed three arguments, the Builder object which is assigning a new job, a list of WorkerForBuilder objects and the BuildRequest. The function should return one of the WorkerForBuilder objects, or None if none of the available workers should be used. As an example, for each worker in the list, worker.worker will be a Worker object, and worker.worker.workername is the worker's name. The function can optionally return a Deferred, which should fire with the same results.
If provided, this is a function that controls which build request will be handled next. The function is passed two arguments, the Builder object which is assigning a new job, and a list of BuildRequest objects of pending builds. The function should return one of the BuildRequest objects, or None if none of the pending builds should be started. This function can optionally return a Deferred which should fire with the same results.
If provided, this is a function that can veto whether a particular worker should be used for a given build request. The function is passed three arguments: the Builder, a Worker, and a BuildRequest. The function should return True if the combination is acceptable, or False otherwise. This function can optionally return a Deferred which should fire with the same results.
This argument specifies a list of locks that apply to this builder; see Interlocks.

A Builder may be given a dictionary of environment variables in this parameter. The variables are used in ShellCommand steps in builds created by this builder. The environment variables will override anything in the worker's environment. Variables passed directly to a ShellCommand will override variables of the same name passed to the Builder.

For example, if you have a pool of identical workers it is often easier to manage variables like PATH from Buildbot rather than manually editing it inside of the workers' environment.

f = factory.BuildFactory
              command=['bash', './configure']))

c['builders'] = [
  BuilderConfig(name='test', factory=f,
        workernames=['worker1', 'worker2', 'worker3', 'worker4'],
        env={'PATH': '/opt/local/bin:/opt/app/bin:/usr/local/bin:/usr/bin'}),

Unlike most builder configuration arguments, this argument can contain renderables.

Specifies how build requests for this builder should be collapsed. See Collapsing Build Requests, below.
A builder may be given a dictionary of Build Properties specific for this builder in this parameter. Those values can be used later on like other properties. Interpolate.
A builder may be given an arbitrary description, which will show up in the web status on the builder's page. Collapsing Build Requests

When more than one build request is available for a builder, Buildbot can "collapse" the requests into a single build. This is desirable when build requests arrive more quickly than the available workers can satisfy them, but has the drawback that separate results for each build are not available.

Requests are only candidated for a merge if both requests have exactly the same codebases.

This behavior can be controlled globally, using the collapseRequests parameter, and on a per-Builder basis, using the collapseRequests argument to the Builder configuration. If collapseRequests is given, it completely overrides the global configuration.

Possible values for both collapseRequests configurations are:

Requests will be collapsed if their sourcestamp are compatible (see below for definition of compatible).
Requests will never be collapsed.
callable(builder, req1, req2)
Requests will be collapsed if the callable returns true. See Collapse Request Functions for detailed example.

Sourcestamps are compatible if all of the below conditions are met:

  • Their codebase, branch, project, and repository attributes match exactly
  • Neither source stamp has a patch (e.g., from a try scheduler)
  • Either both source stamps are associated with changes, or neither are associated with changes but they have matching revisions. Prioritizing Builds

The BuilderConfig parameter nextBuild can be use to prioritize build requests within a builder. Note that this is orthogonal to Prioritizing Builders, which controls the order in which builders are called on to start their builds. The details of writing such a function are in Build Priority Functions.

Such a function can be provided to the BuilderConfig as follows:

def pickNextBuild(builder, requests):
c['builders'] = [
    BuilderConfig(name='test', factory=f,
        workernames=['worker1', 'worker2', 'worker3', 'worker4']),

2.4.7. Build Factories

Each Builder is equipped with a build factory, which defines the steps used to perform that particular type of build. This factory is created in the configuration file, and attached to a Builder through the factory element of its dictionary.

The steps used by these builds are defined in the next section, Build Steps.


Build factories are used with builders, and are not added directly to the buildmaster configuration dictionary. Defining a Build Factory

A BuildFactory defines the steps that every build will follow. Think of it as a glorified script. For example, a build factory which consists of an SVN checkout followed by a make build would be configured as follows:

from buildbot.plugins import util, steps

f = util.BuildFactory()
f.addStep(steps.SVN(repourl="http://..", mode="incremental"))
f.addStep(steps.Compile(command=["make", "build"]))

This factory would then be attached to one builder (or several, if desired):

    BuilderConfig(name='quick', workernames=['bot1', 'bot2'], factory=f))

It is also possible to pass a list of steps into the BuildFactory when it is created. Using addStep is usually simpler, but there are cases where it is more convenient to create the list of steps ahead of time, perhaps using some Python tricks to generate the steps.

from buildbot.plugins import steps, util

all_steps = [
    steps.CVS(cvsroot=CVSROOT, cvsmodule="project", mode="update"),
    steps.Compile(command=["make", "build"]),
f = util.BuildFactory(all_steps)

Finally, you can also add a sequence of steps all at once:


The following attributes can be set on a build factory after it is created, e.g.,

f = util.BuildFactory()
f.useProgress = False
(defaults to True): if True, the buildmaster keeps track of how long each step takes, so it can provide estimates of how long future builds will take. If builds are not expected to take a consistent amount of time (such as incremental builds in which a random set of files are recompiled or tested each time), this should be set to False to inhibit progress-tracking.

(defaults to 'build'): workdir given to every build step created by this factory as default. The workdir can be overridden in a build step definition.

If this attribute is set to a string, that string will be used for constructing the workdir (worker base + builder builddir + workdir). The attribute can also be a Python callable, for more complex cases, as described in Factory Workdir Functions. Dynamic Build Factories

In some cases you may not know what commands to run until after you checkout the source tree. For those cases you can dynamically add steps during a build from other steps.

The Build object provides 2 functions to do this:

addStepsAfterCurrentStep(self, step_factories)
This adds the steps after the step that is currently executing.
addStepsAfterLastStep(self, step_factories)
This adds the steps onto the end of the build.

Both functions only accept as an argument a list of steps to add to the build.

For example lets say you have a script checked in into your source tree called When this script is called with the argument --list-stages it outputs a newline separated list of stage names. This can be used to generate at runtime a step for each stage in the build. Each stage is then run in this example using ./ --run-stage <stage name>.

from buildbot.plugins import util, steps
from buildbot.process import buildstep, logobserver
from twisted.internet import defer

class GenerateStagesCommand(buildstep.ShellMixin, steps.BuildStep):

    def __init__(self, **kwargs):
        kwargs = self.setupShellMixin(kwargs)
        steps.BuildStep.__init__(self, **kwargs) = logobserver.BufferLogObserver()

    def extract_stages(self, stdout):
        stages = []
        for line in stdout.split('\n'):
            stage = str(line.strip())
            if stage:
        return stages

    def run(self):
        # run './ --list-stages' to generate the list of stages
        cmd = yield self.makeRemoteShellCommand()
        yield self.runCommand(cmd)

        # if the command passes extract the list of stages
        result = cmd.results()
        if result == util.SUCCESS:
            # create a ShellCommand for each stage and add them to the build
                steps.ShellCommand(name=stage, command=["./", "--run-stage", stage])
                for stage in self.extract_stages(


f = util.BuildFactory()
    name="Generate build stages",
    command=["./", "--list-stages"],
    haltOnFailure=True)) Predefined Build Factories

Buildbot includes a few predefined build factories that perform common build sequences. In practice, these are rarely used, as every site has slightly different requirements, but the source for these factories may provide examples for implementation of those requirements.

class buildbot.process.factory.GNUAutoconf

GNU Autoconf is a software portability tool, intended to make it possible to write programs in C (and other languages) which will run on a variety of UNIX-like systems. Most GNU software is built using autoconf. It is frequently used in combination with GNU automake. These tools both encourage a build process which usually looks like this:

% CONFIG_ENV=foo ./configure --with-flags
% make all
% make check
# make install

(except of course the Buildbot always skips the make install part).

The Buildbot's buildbot.process.factory.GNUAutoconf factory is designed to build projects which use GNU autoconf and/or automake. The configuration environment variables, the configure flags, and command lines used for the compile and test are all configurable, in general the default values will be suitable.


f = util.GNUAutoconf(source=source.SVN(repourl=URL, mode="copy"),

Required Arguments:

This argument must be a step specification tuple that provides a BuildStep to generate the source tree.

Optional Arguments:

The command used to configure the tree. Defaults to ./configure. Accepts either a string or a list of shell argv elements.
The environment used for the initial configuration step. This accepts a dictionary which will be merged into the worker's normal environment. This is commonly used to provide things like CFLAGS="-O2 -g" (to turn off debug symbols during the compile). Defaults to an empty dictionary.
A list of flags to be appended to the argument list of the configure command. This is commonly used to enable or disable specific features of the autoconf-controlled package, like ["--without-x"] to disable windowing support. Defaults to an empty list.
use autoreconf to generate the ./configure file, set to True to use a buildbot default autoreconf command, or define the command for the ShellCommand.
this is a shell command or list of argv values which is used to actually compile the tree. It defaults to make all. If set to None, the compile step is skipped.
this is a shell command or list of argv values which is used to run the tree's self-tests. It defaults to make check. If set to None, the test step is skipped.
this is a shell command or list of argv values which is used to run the packaging test. It defaults to make distcheck. If set to None, the test step is skipped.
class buildbot.process.factory.BasicBuildFactory

This is a subclass of GNUAutoconf which assumes the source is in CVS, and uses mode='full' and method='clobber' to always build from a clean working copy.

class buildbot.process.factory.BasicSVN

This class is similar to QuickBuildFactory, but uses SVN instead of CVS.

class buildbot.process.factory.QuickBuildFactory

The QuickBuildFactory class is a subclass of GNUAutoconf which assumes the source is in CVS, and uses mode='incremental' to get incremental updates.

The difference between a full build and a quick build is that quick builds are generally done incrementally, starting with the tree where the previous build was performed. That simply means that the source-checkout step should be given a mode='incremental' flag, to do the source update in-place.

In addition to that, this class sets the useProgress flag to False. Incremental builds will (or at least the ought to) compile as few files as necessary, so they will take an unpredictable amount of time to run. Therefore it would be misleading to claim to predict how long the build will take.

This class is probably not of use to new projects.

class buildbot.process.factory.CPAN

Most Perl modules available from the CPAN archive use the MakeMaker module to provide configuration, build, and test services. The standard build routine for these modules looks like:

% perl Makefile.PL
% make
% make test
# make install

(except again Buildbot skips the install step)

Buildbot provides a CPAN factory to compile and test these projects.


(required): A step specification tuple, like that used by GNUAutoconf.
A string which specifies the perl executable to use. Defaults to just perl.
class buildbot.process.factory.Distutils

Most Python modules use the distutils package to provide configuration and build services. The standard build process looks like:

% python ./ build
% python ./ install

Unfortunately, although Python provides a standard unit-test framework named unittest, to the best of my knowledge distutils does not provide a standardized target to run such unit tests. (Please let me know if I'm wrong, and I will update this factory.)

The Distutils factory provides support for running the build part of this process. It accepts the same source= parameter as the other build factories.


(required): A step specification tuple, like that used by GNUAutoconf.
A string which specifies the python executable to use. Defaults to just python.
Provides a shell command which runs unit tests. This accepts either a string or a list. The default value is None, which disables the test step (since there is no common default command to run unit tests in distutils modules).
class buildbot.process.factory.Trial

Twisted provides a unit test tool named trial which provides a few improvements over Python's built-in unittest module. Many Python projects which use Twisted for their networking or application services also use trial for their unit tests. These modules are usually built and tested with something like the following:

% python ./ build
% PYTHONPATH=build/lib.linux-i686-2.3 trial -v PROJECTNAME.test
% python ./ install

Unfortunately, the build/lib directory into which the built/copied .py files are placed is actually architecture-dependent, and I do not yet know of a simple way to calculate its value. For many projects it is sufficient to import their libraries in place from the tree's base directory (PYTHONPATH=.).

In addition, the PROJECTNAME value where the test files are located is project-dependent: it is usually just the project's top-level library directory, as common practice suggests the unit test files are put in the test sub-module. This value cannot be guessed, the Trial class must be told where to find the test files.

The Trial class provides support for building and testing projects which use distutils and trial. If the test module name is specified, trial will be invoked. The library path used for testing can also be set.

One advantage of trial is that the Buildbot happens to know how to parse trial output, letting it identify which tests passed and which ones failed. The Buildbot can then provide fine-grained reports about how many tests have failed, when individual tests fail when they had been passing previously, etc.

Another feature of trial is that you can give it a series of source .py files, and it will search them for special test-case-name tags that indicate which test cases provide coverage for that file. Trial can then run just the appropriate tests. This is useful for quick builds, where you want to only run the test cases that cover the changed functionality.


Provides a directory to add to PYTHONPATH when running the unit tests, if tests are being run. Defaults to . to include the project files in-place. The generated build library is frequently architecture-dependent, but may simply be build/lib for pure-Python modules.
which Python executable to use. This list will form the start of the argv array that will launch trial. If you use this, you should set trial to an explicit path (like /usr/bin/trial or ./bin/trial). The parameter defaults to None, which leaves it out entirely (running trial args instead of python ./bin/trial args). Likely values are ['python'], ['python2.2'], or ['python', '-Wall'].
provides the name of the trial command. It is occasionally useful to use an alternate executable, such as trial2.2 which might run the tests under an older version of Python. Defaults to trial.
a list of arguments to pass to trial, specifically to set the reporting mode. This defaults to ['--reporter=bwverbose'], which only works for Twisted-2.1.0 and later.
a list of arguments to pass to trial, available to turn on any extra flags you like. Defaults to [].
Provides a module name or names which contain the unit tests for this project. Accepts a string, typically PROJECTNAME.test, or a list of strings. Defaults to None, indicating that no tests should be run. You must either set this or testChanges.
if True, ignore the tests parameter and instead ask the Build for all the files that make up the Changes going into this build. Pass these filenames to trial and ask it to look for test-case-name tags, running just the tests necessary to cover the changes.
If True, tells Trial (with the --recurse argument) to look in all subdirectories for additional test cases.
which reactor to use, like 'gtk' or 'java'. If not provided, the Twisted's usual platform-dependent default is used.
If True, tells Trial (with the --random=0 argument) to run the test cases in random order, which sometimes catches subtle inter-test dependency bugs. Defaults to False.

The step can also take any of the ShellCommand arguments, e.g., haltOnFailure.

Unless one of tests or testChanges are set, the step will generate an exception.

2.4.8. Properties

Build properties are a generalized way to provide configuration information to build steps; see Build Properties for the conceptual overview of properties.

Some build properties come from external sources and are set before the build begins; others are set during the build, and available for later steps. The sources for properties are:

global configuration
These properties apply to all builds.
A scheduler can specify properties that become available to all builds it starts.
A change can have properties attached to it, supplying extra information gathered by the change source. This is most commonly used with the sendchange command.
forced builds
The "Force Build" form allows users to specify properties
A worker can pass properties on to the builds it performs.
A build automatically sets a number of properties on itself.
A builder can set properties on all the builds it runs.
The steps of a build can set properties that are available to subsequent steps. In particular, source steps set the got_revision property.

If the same property is supplied in multiple places, the final appearance takes precedence. For example, a property set in a builder configuration will override one supplied by a scheduler.

Properties are stored internally in JSON format, so they are limited to basic types of data: numbers, strings, lists, and dictionaries. Common Build Properties

The following build properties are set when the build is started, and are available to all steps.


This property is set when a Source step checks out the source tree, and provides the revision that was actually obtained from the VC system. In general this should be the same as revision, except for non-absolute sourcestamps, where got_revision indicates what revision was current when the checkout was performed. This can be used to rebuild the same source code later.


For some VC systems (Darcs in particular), the revision is a large string containing newlines, and is not suitable for interpolation into a filename.

For multi-codebase builds (where codebase is not the default ''), this property is a dictionary, keyed by codebase.

This is a string that indicates which Builder the build was a part of. The combination of buildername and buildnumber uniquely identify a build.
Each build gets a number, scoped to the Builder (so the first build performed on any given Builder will have a build number of 0). This integer property contains the build's number.
This is a string which identifies which worker the build is running on.
If the build was started from a scheduler, then this property will contain the name of that scheduler.
The absolute path of the base working directory on the worker, of the current builder.

For single codebase builds, where the codebase is '', the following Source Stamp Attributes are also available as properties: branch, revision, repository, and project . Source Stamp Attributes

branch revision repository project codebase

For details of these attributes see Concepts.


This attribute is a list of dictionaries representing the changes that make up this sourcestamp. Using Properties in Steps

For the most part, properties are used to alter the behavior of build steps during a build. This is done by using renderables (objects implementing the IRenderable interface) as step parameters. When the step is started, each such object is rendered using the current values of the build properties, and the resultant rendering is substituted as the actual value of the step parameter.

Buildbot offers several renderable object types covering common cases. It's also possible to create custom renderables.


Properties are defined while a build is in progress; their values are not available when the configuration file is parsed. This can sometimes confuse newcomers to Buildbot! In particular, the following is a common error:

if Property('release_train') == 'alpha':

This does not work because the value of the property is not available when the if statement is executed. However, Python will not detect this as an error - you will just never see the step added to the factory.

You can use renderables in most step parameters. Please file bugs for any parameters which do not accept renderables.


The simplest renderable is Property, which renders to the value of the property named by its argument:

from buildbot.plugins import steps, util

f.addStep(steps.ShellCommand(command=['echo', 'buildername:',

You can specify a default value by passing a default keyword argument:

f.addStep(steps.ShellCommand(command=['echo', 'warnings:',
                             util.Property('warnings', default='none')]))

The default value is used when the property doesn't exist, or when the value is something Python regards as False. The defaultWhenFalse argument can be set to False to force buildbot to use the default argument only if the parameter is not set:

f.addStep(steps.ShellCommand(command=['echo', 'warnings:',
                             util.Property('warnings', default='none',

The default value can be a renderable itself, e.g.,

command=util.Property('command', default=util.Property('default-command'))

Property can only be used to replace an entire argument: in the example above, it replaces an argument to echo. Often, properties need to be interpolated into strings, instead. The tool for that job is Interpolate.

The more common pattern is to use Python dictionary-style string interpolation by using the %(prop:<propname>)s syntax. In this form, the property name goes in the parentheses, as above. A common mistake is to omit the trailing "s", leading to a rather obscure error from Python ("ValueError: unsupported format character").

from buildbot.plugins import steps, util

This example will result in a make command with an argument like REVISION=12098.

The syntax of dictionary-style interpolation is a selector, followed by a colon, followed by a selector specific key, optionally followed by a colon and a string indicating how to interpret the value produced by the key.

The following selectors are supported.

The key is the name of a property.
The key is a codebase and source stamp attribute, separated by a colon.
The key refers to a keyword argument passed to Interpolate. Those keyword arguments may be ordinary values or renderables.

The following ways of interpreting the value are available.

If the key exists, substitute its value; otherwise, substitute replacement. replacement may be empty (%(prop:propname:-)s). This is the default.
Like -replacement, but only substitutes the value of the key if it is something Python regards as True. Python considers None, 0, empty lists, and the empty string to be false, so such values will be replaced by replacement.
If the key exists, substitute replacement; otherwise, substitute an empty string.


Ternary substitution, depending on either the key being present (with ?, similar to +) or being True (with #?, like ~). Notice that there is a pipe immediately following the question mark and between the two substitution alternatives. The character that follows the question mark is used as the delimiter between the two alternatives. In the above examples, it is a pipe, but any character other than ( can be used.


Although these are similar to shell substitutions, no other substitutions are currently supported.


from buildbot.plugins import steps, util

In addition, Interpolate supports using positional string interpolation. Here, %s is used as a placeholder, and the substitutions (which may be renderables), are given as subsequent arguments:



Like Python, you can use either positional interpolation or dictionary-style interpolation, not both. Thus you cannot use a string like Interpolate("foo-%(src::revision)s-%s", "branch").


While Interpolate can handle many simple cases, and even some common conditionals, more complex cases are best handled with Python code. The renderer decorator creates a renderable object whose rendering is obtained by calling the decorated function when the step it's passed to begins. The function receives an IProperties object, which it can use to examine the values of any and all properties. For example:

from buildbot.plugins import steps, util

def makeCommand(props):
    command = ['make']
    cpus = props.getProperty('CPUs')
    if cpus:
        command.extend(['-j', str(cpus+1)])
        command.extend(['-j', '2'])
    return command


You can think of renderer as saying "call this function when the step starts".

Note: Config errors with Renderables may not always be caught via checkconfig


Transform is an alternative to renderer. While renderer is useful for creating new renderables, Transform is easier to use when you want to transform or combine the renderings of preexisting ones.

Transform takes a function and any number of positional and keyword arguments. The function must either be a callable object or a renderable producing one. When rendered, a Transform first replaces all of its arguments that are renderables with their renderings, then calls the function, passing it the positional and keyword arguments, and returns the result as its own rendering.

For example, suppose my_path is a path on the worker, and you want to get it relative to the build directory. You can do it like this:

import os.path
from buildbot.plugins import util

my_path_rel = util.Transform(os.path.relpath, my_path, start=util.Property('builddir'))

This works whether my_path is an ordinary string or a renderable. my_path_rel will be a renderable in either case, however.


If nested list should be flatten for some renderables, FlattenList could be used. For example:

f.addStep(ShellCommand(command=[ 'make' ], descriptionDone=FlattenList([ 'make ', [ 'done' ]])))

descriptionDone would be set to [ 'make', 'done' ] when the ShellCommand executes. This is useful when a list-returning property is used in renderables.


ShellCommand automatically flattens nested lists in its command argument, so there is no need to use FlattenList for it.



This class is deprecated. It is an older version of Interpolate. It exists for compatibility with older configs.

The simplest use of this class is with positional string interpolation. Here, %s is used as a placeholder, and property names are given as subsequent arguments:

from buildbot.plugins import steps, util
    command=["tar", "czf",
             util.WithProperties("build-%s-%s.tar.gz", "branch", "revision"),

If this BuildStep were used in a tree obtained from Git, it would create a tarball with a name like build-master-a7d3a333db708e786edb34b6af646edd8d4d3ad9.tar.gz.

The more common pattern is to use Python dictionary-style string interpolation by using the %(propname)s syntax. In this form, the property name goes in the parentheses, as above. A common mistake is to omit the trailing "s", leading to a rather obscure error from Python ("ValueError: unsupported format character").

from buildbot.plugins import steps, util

This example will result in a make command with an argument like REVISION=12098.

The dictionary-style interpolation supports a number of more advanced syntaxes in the parentheses.

If propname exists, substitute its value; otherwise, substitute replacement. replacement may be empty (%(propname:-)s)
Like propname:-replacement, but only substitutes the value of property propname if it is something Python regards as True. Python considers None, 0, empty lists, and the empty string to be false, so such values will be replaced by replacement.
If propname exists, substitute replacement; otherwise, substitute an empty string.

Although these are similar to shell substitutions, no other substitutions are currently supported, and replacement in the above cannot contain more substitutions.

Note: like Python, you can use either positional interpolation or dictionary-style interpolation, not both. Thus you cannot use a string like WithProperties("foo-%(revision)s-%s", "branch").

Custom Renderables

If the options described above are not sufficient, more complex substitutions can be achieved by writing custom renderables.

The IRenderable interface is simple - objects must provide a getRenderingFor method. The method should take one argument - an IProperties provider - and should return the rendered value or a deferred firing with one. Pass instances of the class anywhere other renderables are accepted. For example:

class DetermineFoo(object):
    def getRenderingFor(self, props):
        if props.hasProperty('bar'):
            return props['bar']
        elif props.hasProperty('baz'):
            return props['baz']
        return 'qux'
ShellCommand(command=['echo', DetermineFoo()])

or, more practically,

class Now(object):
    def getRenderingFor(self, props):
        return time.clock()
ShellCommand(command=['make', Interpolate('TIME=%(kw:now)s', now=Now())])

This is equivalent to:

def now(props):
    return time.clock()
ShellCommand(command=['make', Interpolate('TIME=%(kw:now)s', now=now)])

Note that a custom renderable must be instantiated (and its constructor can take whatever arguments you'd like), whereas a function decorated with renderer can be used directly.

URL for build

Its common to need to use the URL for the build in a step. For this you can use a special custom renderer as following:

from buildbot.plugins import *

ShellCommand(command=['make', Interpolate('BUILDURL=%(kw:url)s', url=util.URLForBuild)])

2.4.9. Build Steps

BuildSteps are usually specified in the buildmaster's configuration file, in a list that goes into the BuildFactory. The BuildStep instances in this list are used as templates to construct new independent copies for each build (so that state can be kept on the BuildStep in one build without affecting a later build). Each BuildFactory can be created with a list of steps, or the factory can be created empty and then steps added to it using the addStep method:

from buildbot.plugins import util, steps

f = util.BuildFactory()
    steps.ShellCommand(command=["make", "all"]),
    steps.ShellCommand(command=["make", "test"])

The basic behavior for a BuildStep is to:

  • run for a while, then stop
  • possibly invoke some RemoteCommands on the attached worker
  • possibly produce a set of log files
  • finish with a status described by one of four values defined in buildbot.status.builder: SUCCESS, WARNINGS, FAILURE, SKIPPED
  • provide a list of short strings to describe the step

The rest of this section describes all the standard BuildStep objects available for use in a Build, and the parameters which can be used to control each. A full list of build steps is available in the Build Step Index. Common Parameters

All BuildSteps accept some common parameters. Some of these control how their individual status affects the overall build. Others are used to specify which Locks (see Interlocks) should be acquired before allowing the step to run.

Arguments common to all BuildStep subclasses:

the name used to describe the step on the status display. It is also used to give a name to any LogFiles created by this step.
if True, a FAILURE of this build step will cause the build to halt immediately. Steps with alwaysRun=True are still run. Generally speaking, haltOnFailure implies flunkOnFailure (the default for most BuildSteps). In some cases, particularly series of tests, it makes sense to haltOnFailure if something fails early on but not flunkOnFailure. This can be achieved with haltOnFailure=True, flunkOnFailure=False.
when True, a WARNINGS or FAILURE of this build step will mark the overall build as FAILURE. The remaining steps will still be executed.
when True, a FAILURE of this build step will mark the overall build as a FAILURE. The remaining steps will still be executed.
when True, a WARNINGS or FAILURE of this build step will mark the overall build as having WARNINGS. The remaining steps will still be executed.
when True, a FAILURE of this build step will mark the overall build as having WARNINGS. The remaining steps will still be executed.
if True, this build step will always be run, even if a previous buildstep with haltOnFailure=True has failed.
This will be used to describe the command (on the Waterfall display) while the command is still running. It should be a single imperfect-tense verb, like compiling or testing. The preferred form is a single, short string, but for historical reasons a list of strings is also acceptable.

This will be used to describe the command once it has finished. A simple noun like compile or tests should be used. Like description, this may either be a string or a list of short strings.

If neither description nor descriptionDone are set, the actual command arguments will be used to construct the description. This may be a bit too wide to fit comfortably on the Waterfall display.

All subclasses of BuildStep will contain the description attributes. Consequently, you could add a ShellCommand step like so:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "test"],

This is an optional suffix appended to the end of the description (ie, after description and descriptionDone). This can be used to distinguish between build steps that would display the same descriptions in the waterfall. This parameter may be a string, a list of short strings or None.

For example, a builder might use the Compile step to build two different codebases. The descriptionSuffix could be set to projectFoo and projectBar, respectively for each step, which will result in the full descriptions compiling projectFoo and compiling projectBar to be shown in the waterfall.

A step can be configured to only run under certain conditions. To do this, set the step's doStepIf to a boolean value, or to a function that returns a boolean value or Deferred. If the value or function result is false, then the step will return SKIPPED without doing anything. Otherwise, the step will be executed normally. If you set doStepIf to a function, that function should accept one parameter, which will be the Step object itself.

A step can be optionally hidden from the waterfall and build details web pages. To do this, set the step's hideStepIf to a boolean value, or to a function that takes two parameters -- the results and the BuildStep -- and returns a boolean value. Steps are always shown while they execute, however after the step as finished, this parameter is evaluated (if a function) and if the value is True, the step is hidden. For example, in order to hide the step if the step has been skipped:

factory.addStep(Foo(..., hideStepIf=lambda results, s: results==SKIPPED))
a list of Locks (instances of buildbot.locks.WorkerLock or buildbot.locks.MasterLock) that should be acquired before starting this BuildStep. Alternatively this could be a renderable that returns this list during build execution. This lets you defer picking the locks to acquire until the build step is about to start running. The Locks will be released when the step is complete. Note that this is a list of actual Lock instances, not names. Also note that all Locks must have unique names. See Interlocks.
The character encoding to use to decode logs produced during the execution of this step. This overrides the default logEncoding; see Log Handling. Source Checkout


Support for the old worker-side source checkout steps was removed in Buildbot-0.9.0.

The old source steps used to be imported like this:

from buildbot.steps.source.oldsource import Git

... Git ...


from buildbot.steps.source import Git

while new source steps are in separate Python modules for each version-control system and, using the plugin infrastructure are available as:

from buildbot.plugins import steps

... steps.Git ...
Common Parameters

All source checkout steps accept some common parameters to control how they get the sources and where they should be placed. The remaining per-VC-system parameters are mostly to specify where exactly the sources are coming from.

mode method

These two parameters specify the means by which the source is checked out. mode specifies the type of checkout and method tells about the way to implement it.

from buildbot.plugins import steps

factory = BuildFactory()
factory.addStep(steps.Mercurial(repourl='path/to/repo', mode='full',

The mode parameter a string describing the kind of VC operation that is desired, defaulting to incremental. The options are

Update the source to the desired revision, but do not remove any other files generated by previous builds. This allows compilers to take advantage of object files from previous builds. This mode is exactly same as the old update mode.
Update the source, but delete remnants of previous builds. Build steps that follow will need to regenerate all object files.

Methods are specific to the version-control system in question, as they may take advantage of special behaviors in that version-control system that can make checkouts more efficient or reliable.

like all Steps, this indicates the directory where the build will take place. Source Steps are special in that they perform some operations outside of the workdir (like creating the workdir itself).
if True, bypass the usual behavior of checking out the revision in the source stamp, and always update to the latest revision in the repository instead.
If set, this specifies a tuple of (delay, repeats) which means that when a full VC checkout fails, it should be retried up to repeats times, waiting delay seconds between attempts. If you don't provide this, it defaults to None, which means VC operations should not be retried. This is provided to make life easier for workers which are stuck behind poor network connections.

The name of this parameter might vary depending on the Source step you are running. The concept explained here is common to all steps and applies to repourl as well as for baseURL (when applicable).

A common idiom is to pass Property('repository', 'url://default/repo/path') as repository. This grabs the repository from the source stamp of the build. This can be a security issue, if you allow force builds from the web, or have the WebStatus change hooks enabled; as the worker will download code from an arbitrary repository.

This specifies which codebase the source step should use to select the right source stamp. The default codebase value is ''. The codebase must correspond to a codebase assigned by the codebaseGenerator. If there is no codebaseGenerator defined in the master then codebase doesn't need to be set, the default value will then match all changes.
Specifies the timeout for worker-side operations, in seconds. If your repositories are particularly large, then you may need to increase this value from its default of 1200 (20 minutes).
If this option is true (the default), then the step's logfile will describe the environment variables on the worker. In situations where the environment is not relevant and is long, it may be easier to set logEnviron=False.
a dictionary of environment strings which will be added to the child command's environment. The usual property interpolations can be used in environment variable names and values - see Properties.
class buildbot.steps.source.mercurial.Mercurial

The Mercurial build step performs a Mercurial (aka hg) checkout or update.

Branches are available in two modes: dirname, where the name of the branch is a suffix of the name of the repository, or inrepo, which uses Hg's named-branches support. Make sure this setting matches your changehook, if you have that installed.

from buildbot.plugins import steps

factory.addStep(steps.Mercurial(repourl='path/to/repo', mode='full',
                                method='fresh', branchType='inrepo'))

The Mercurial step takes the following arguments:

where the Mercurial source repository is available.
this specifies the name of the branch to use when a Build does not provide one of its own. This will be appended to repourl to create the string that will be passed to the hg clone command.
either 'dirname' (default) or 'inrepo' depending on whether the branch name should be appended to the repourl or the branch is a Mercurial named branch and can be found within the repourl.
boolean, defaults to True. If set and using inrepos branches, clobber the tree at each branch change. Otherwise, just update to the branch.

mode method

Mercurial's incremental mode does not require a method. The full mode has three methods defined:

It removes the build directory entirely then makes full clone from repo. This can be slow as it need to clone whole repository
This remove all other files except those tracked by VCS. First it does hg purge --all then pull/update
All the files which are tracked by Mercurial and listed ignore files are not deleted. Remaining all other files will be deleted before pull/update. This is equivalent to hg purge then pull/update.
class buildbot.steps.source.git.Git

The Git build step clones or updates a Git repository and checks out the specified branch or revision.


The Buildbot supports Git version 1.2.0 and later: earlier versions (such as the one shipped in Ubuntu 'Dapper') do not support the git init command that the Buildbot uses.

from buildbot.plugins import steps

factory.addStep(steps.Git(repourl='git://path/to/repo', mode='full',
                          method='clobber', submodules=True))

The Git step takes the following arguments:

(required): the URL of the upstream Git repository.
(optional): this specifies the name of the branch to use when a Build does not provide one of its own. If this this parameter is not specified, and the Build does not provide a branch, the default branch of the remote repository will be used.
(optional): when initializing/updating a Git repository, this tells Buildbot whether to handle Git submodules. Default: False.
(optional): instructs git to attempt shallow clones (--depth 1). The depth defaults to 1 and can be changed by passing an integer instead of True. This option can be used only in full builds with clobber method.
(optional): use the specified string as a path to a reference repository on the local machine. Git will try to grab objects from this path first instead of the main repository, if they exist.
(optional): By default, any clone will use the name "origin" as the remote repository (eg, "origin/master"). This renderable option allows that to be configured to an alternate name.
(optional): passes the (--progress) flag to (git fetch). This solves issues of long fetches being killed due to lack of output, but requires Git 1.7.2 or later.
(optional): defaults to False. If true, if the git fetch fails then buildbot retries to fetch again instead of failing the entire source checkout.
(optional): defaults to False. If a fetch or full clone fails we can checkout source removing everything. This way new repository will be cloned. If retry fails it fails the source checkout step.


(optional): defaults to 'incremental'. Specifies whether to clean the build tree or not.

The source is update, but any built files are left untouched.
The build tree is clean of any built files. The exact method for doing this is controlled by the method argument.


(optional): defaults to fresh when mode is full. Git's incremental mode does not require a method. The full mode has four methods defined:

It removes the build directory entirely then makes full clone from repo. This can be slow as it need to clone whole repository. To make faster clones enable shallow option. If shallow options is enabled and build request have unknown revision value, then this step fails.
This remove all other files except those tracked by Git. First it does git clean -d -f -f -x then fetch/checkout to a specified revision(if any). This option is equal to update mode with ignore_ignores=True in old steps.
All the files which are tracked by Git and listed ignore files are not deleted. Remaining all other files will be deleted before fetch/checkout. This is equivalent to git clean -d -f -f then fetch. This is equivalent to ignore_ignores=False in old steps.
This first checkout source into source directory then copy the source directory to build directory then performs the build operation in the copied directory. This way we make fresh builds with very less bandwidth to download source. The behavior of source checkout follows exactly same as incremental. It performs all the incremental checkout behavior in source directory.


(optional) After checkout, invoke a git describe on the revision and save the result in a property; the property's name is either commit-description or commit-description-foo, depending on whether the codebase argument was also provided. The argument should either be a bool or dict, and will change how git describe is called:

  • getDescription=False: disables this feature explicitly

  • getDescription=True or empty dict(): Run git describe with no args

  • getDescription={...}: a dict with keys named the same as the Git option. Each key's value can be False or None to explicitly skip that argument.

    For the following keys, a value of True appends the same-named Git argument:

    • all : --all
    • always: --always
    • contains: --contains
    • debug: --debug
    • long: --long`
    • exact-match: --exact-match
    • tags: --tags
    • dirty: --dirty

    For the following keys, an integer or string value (depending on what Git expects) will set the argument's parameter appropriately. Examples show the key-value pair:

    • match=foo: --match foo
    • abbrev=7: --abbrev=7
    • candidates=7: --candidates=7
    • dirty=foo: --dirty=foo


(optional) A dict of git configuration settings to pass to the remote git commands.
class buildbot.steps.source.svn.SVN

The SVN build step performs a Subversion checkout or update. There are two basic ways of setting up the checkout step, depending upon whether you are using multiple branches or not.

The SVN step should be created with the repourl argument:

(required): this specifies the URL argument that will be given to the svn checkout command. It dictates both where the repository is located and which sub-tree should be extracted. One way to specify the branch is to use Interpolate. For example, if you wanted to check out the trunk repository, you could use repourl=Interpolate(""). Alternatively, if you are using a remote Subversion repository which is accessible through HTTP at a URL of, and you wanted to check out the trunk/calc sub-tree, you would directly use repourl="" as an argument to your SVN step.

If you are building from multiple branches, then you should create the SVN step with the repourl and provide branch information with Interpolate:

from buildbot.plugins import steps, util


Alternatively, the repourl argument can be used to create the SVN step without Interpolate:

from buildbot.plugins import steps

(optional): if specified, this will be passed to the svn binary with a --username option.
(optional): if specified, this will be passed to the svn binary with a --password option.
(optional): if specified, an array of strings that will be passed as extra arguments to the svn binary.
(optional): specific files or directories to keep between purges, like some build outputs that can be reused between builds.

(optional): Specify depth argument to achieve sparse checkout. Only available if worker has Subversion 1.5 or higher.

If set to empty updates will not pull in any files or subdirectories not already present. If set to files, updates will pull in any files not already present, but not directories. If set to immediates, updates will pull in any files or subdirectories not already present, the new subdirectories will have depth: empty. If set to infinity, updates will pull in any files or subdirectories not already present; the new subdirectories will have depth-infinity. Infinity is equivalent to SVN default update behavior, without specifying any depth argument.

(optional): By default, the got_revision property is set to the repository's global revision ("Revision" in the svn info output). Set this parameter to True to have it set to the "Last Changed Rev" instead.

mode method

SVN's incremental mode does not require a method. The full mode has five methods defined:

It removes the working directory for each build then makes full checkout.
This always always purges local changes before updating. This deletes unversioned files and reverts everything that would appear in a svn status --no-ignore. This is equivalent to the old update mode with always_purge.
This is same as fresh except that it deletes all unversioned files generated by svn status.
This first checkout source into source directory then copy the source directory to build directory then performs the build operation in the copied directory. This way we make fresh builds with very less bandwidth to download source. The behavior of source checkout follows exactly same as incremental. It performs all the incremental checkout behavior in source directory.
Similar to method='copy', except using svn export to create build directory so that there are no .svn directories in the build directory.

If you are using branches, you must also make sure your ChangeSource will report the correct branch names.

class buildbot.steps.source.cvs.CVS

The CVS build step performs a CVS checkout or update.

from buildbot.plugins import steps


This step takes the following arguments:

(required): specify the CVSROOT value, which points to a CVS repository, probably on a remote machine. For example, if Buildbot was hosted in CVS then the CVSROOT value you would use to get a copy of the Buildbot source code might be
(required): specify the cvs module, which is generally a subdirectory of the CVSROOT. The cvsmodule for the Buildbot source code is buildbot.
a string which will be used in a -r argument. This is most useful for specifying a branch to work on. Defaults to HEAD.
a list of flags to be put before the argument checkout in the CVS command.
a list of flags to be put after the checkout in the CVS command.

mode method

No method is needed for incremental mode. For full mode, method can take the values shown below. If no value is given, it defaults to fresh.

This specifies to remove the workdir and make a full checkout.
This method first runs cvsdisard in the build directory, then updates it. This requires cvsdiscard which is a part of the cvsutil package.
This method is the same as method='fresh', but it runs cvsdiscard --ignore instead of cvsdiscard.
This maintains a source directory for source, which it updates copies to the build directory. This allows Buildbot to start with a fresh directory, without downloading the entire repository on every build.
Password to use while performing login to the remote CVS server. Default is None meaning that no login needs to be peformed.
class buildbot.steps.source.bzr.Bzr

bzr is a descendant of Arch/Baz, and is frequently referred to as simply Bazaar. The repository-vs-workspace model is similar to Darcs, but it uses a strictly linear sequence of revisions (one history per branch) like Arch. Branches are put in subdirectories. This makes it look very much like Mercurial.

from buildbot.plugins import steps


The step takes the following arguments:

(required unless baseURL is provided): the URL at which the Bzr source repository is available.
(required unless repourl is provided): the base repository URL, to which a branch name will be appended. It should probably end in a slash.
(allowed if and only if baseURL is provided): this specifies the name of the branch to use when a Build does not provide one of its own. This will be appended to baseURL to create the string that will be passed to the bzr checkout command.

mode method

No method is needed for incremental mode. For full mode, method can take the values shown below. If no value is given, it defaults to fresh.

This specifies to remove the workdir and make a full checkout.
This method first runs bzr clean-tree to remove all the unversioned files then update the repo. This remove all unversioned files including those in .bzrignore.
This is same as fresh except that it doesn't remove the files mentioned in .bzrginore i.e, by running bzr clean-tree --ignore.
A local bzr repository is maintained and the repo is copied to build directory for each build. Before each build the local bzr repo is updated then copied to build for next steps.
class buildbot.steps.source.p4.P4

The P4 build step creates a Perforce client specification and performs an update.

from buildbot.plugins import steps, util


You can specify the client spec in two different ways. You can use the p4base, p4branch, and (optionally) p4extra_views to build up the viewspec, or you can utilize the p4viewspec to specify the whole viewspec as a set of tuples.

Using p4viewspec will allow you to add lines such as:

//depot/branch/mybranch/...             //<p4client>/...
-//depot/branch/mybranch/notthisdir/... //<p4client>/notthisdir/...

If you specify p4viewspec and any of p4base, p4branch, and/or p4extra_views you will receive a configuration error exception.

A view into the Perforce depot without branch name or trailing /.... Typically //depot/proj.
(optional): A single string, which is appended to the p4base as follows <p4base>/<p4branch>/... to form the first line in the viewspec
(optional): a list of (depotpath, clientpath) tuples containing extra views to be mapped into the client specification. Both will have /... appended automatically. The client name and source directory will be prepended to the client path.

This will override any p4branch, p4base, and/or p4extra_views specified. The viewspec will be an array of tuples as follows:


It yields a viewspec with just:

//depot/main/... //<p4client>/...

(optional): The p4viewspec lets you customize the client spec for a builder but, as the previous example shows, it automatically adds ... at the end of each line. If you need to also specify file-level remappings, you can set the p4viewspec_suffix to None so that nothing is added to your viewspec:

[('//depot/main/...', '...'),
 ('-//depot/main/config.xml', 'config.xml'),
 ('//depot/main/config.vancouver.xml', 'config.xml')]

It yields a viewspec with:

//depot/main/...                  //<p4client>/...
-//depot/main/config.xml          //<p4client/main/config.xml
//depot/main/config.vancouver.xml //<p4client>/main/config.xml

Note how, with p4viewspec_suffix set to None, you need to manually add ... where you need it.

(optional): By default, clients are created with the allwrite rmdir options. This string lets you change that.
(optional): the host:port string describing how to get to the P4 Depot (repository), used as the option -p argument for all p4 commands.
(optional): the Perforce user, used as the option -u argument to all p4 commands.
(optional): the Perforce password, used as the option -p argument to all p4 commands.
(optional): The name of the client to use. In mode='full' and mode='incremental', it's particularly important that a unique name is used for each checkout directory to avoid incorrect synchronization. For this reason, Python percent substitution will be performed on this value to replace %(prop:workername)s with the worker name and %(prop:buildername)s with the builder name. The default is buildbot_%(prop:workername)s_%(prop:buildername)s.
(optional): The type of line ending handling P4 should use. This is added directly to the client spec's LineEnd property. The default is local.

(optional): Extra arguments to be added to the P4 command-line for the sync command. So for instance if you want to sync only to populate a Perforce proxy (without actually syncing files to disk), you can do:

P4(p4extra_args=['-Zproxyload'], ...)
Set to True to use ticket-based authentication, instead of passwords (but you still need to specify p4passwd).
class buildbot.steps.source.repo.Repo

The Repo build step performs a Repo init and sync.

The Repo step takes the following arguments:

(required): the URL at which the Repo's manifests source repository is available.
(optional, defaults to master): the manifest repository branch on which repo will take its manifest. Corresponds to the -b argument to the repo init command.
(optional, defaults to default.xml): the manifest filename. Corresponds to the -m argument to the repo init command.
(optional, defaults to None): the repo tarball used for fast bootstrap. If not present the tarball will be created automatically after first sync. It is a copy of the .repo directory which contains all the Git objects. This feature helps to minimize network usage on very big projects with lots of workers.
(optional, defaults to None): Number of projects to fetch simultaneously while syncing. Passed to repo sync subcommand with "-j".
(optional, defaults to False): renderable boolean to control whether repo syncs all branches. I.e. repo sync -c
(optional, defaults to 0): Depth argument passed to repo init. Specifies the amount of git history to store. A depth of 1 is useful for shallow clones. This can save considerable disk space on very large projects.
(optional, defaults to "one week"): renderable to control the policy of updating of the tarball given properties. Returns: max age of tarball in seconds, or None, if we want to skip tarball update. The default value should be good trade off on size of the tarball, and update frequency compared to cost of tarball creation

(optional, defaults to None): list of repo download commands to perform at the end of the Repo step each string in the list will be prefixed repo download, and run as is. This means you can include parameter in the string. For example:

  • ["-c project 1234/4"] will cherry-pick patchset 4 of patch 1234 in project project
  • ["-f project 1234/4"] will enforce fast-forward on patchset 4 of patch 1234 in project project
class buildbot.steps.source.repo.RepoDownloadsFromProperties

util.repo.DownloadsFromProperties can be used as a renderable of the repoDownload parameter it will look in passed properties for string with following possible format:

  • repo download project change_number/patchset_number
  • project change_number/patchset_number
  • project/change_number/patchset_number

All of these properties will be translated into a repo download. This feature allows integrators to build with several pending interdependent changes, which at the moment cannot be described properly in Gerrit, and can only be described by humans.

class buildbot.steps.source.repo.RepoDownloadsFromChangeSource

util.repo.DownloadsFromChangeSource can be used as a renderable of the repoDownload parameter

This rendereable integrates with GerritChangeSource, and will automatically use the repo download command of repo to download the additionnal changes introduced by a pending changeset.


You can use the two above Rendereable in conjuction by using the class

For example:

from buildbot.plugins import steps, util

class buildbot.steps.source.gerrit.Gerrit

Gerrit step is exactly like the Git step, except that it integrates with GerritChangeSource, and will automatically checkout the additional changes.

Gerrit integration can be also triggered using forced build with property named gerrit_change with values in format change_number/patchset_number. This property will be translated into a branch name. This feature allows integrators to build with several pending interdependent changes, which at the moment cannot be described properly in Gerrit, and can only be described by humans.

class buildbot.steps.source.darcs.Darcs

The Darcs build step performs a Darcs checkout or update.

from buildbot.plugins import steps

                            mode='full', method='clobber', retry=(10, 1)))

Darcs step takes the following arguments:

(required): The URL at which the Darcs source repository is available.


(optional): defaults to 'incremental'. Specifies whether to clean the build tree or not.

The source is update, but any built files are left untouched.
The build tree is clean of any built files. The exact method for doing this is controlled by the method argument.

(optional): defaults to copy when mode is full. Darcs' incremental mode does not require a method. The full mode has two methods defined:

It removes the working directory for each build then makes full checkout.
This first checkout source into source directory then copy the source directory to build directory then performs the build operation in the copied directory. This way we make fresh builds with very less bandwidth to download source. The behavior of source checkout follows exactly same as incremental. It performs all the incremental checkout behavior in source directory.

The Monotone build step performs a Monotone checkout or update.

from buildbot.plugins import steps

                               mode='full', method='clobber',
                               branch='', retry=(10, 1)))

Monotone step takes the following arguments:

the URL at which the Monotone source repository is available.
this specifies the name of the branch to use when a Build does not provide one of its own.
this is a boolean that has a pull from the repository use --ticker=dot instead of the default --ticker=none.


(optional): defaults to 'incremental'. Specifies whether to clean the build tree or not. In any case, the worker first pulls from the given remote repository to synchronize (or possibly initialize) its local database. The mode and method only affect how the build tree is checked-out or updated from the local database.

The source is update, but any built files are left untouched.
The build tree is clean of any built files. The exact method for doing this is controlled by the method argument. Even in this mode, the revisions already pulled remain in the database and a fresh pull is rarely needed.


(optional): defaults to copy when mode is full. Monotone's incremental mode does not require a method. The full mode has four methods defined:

It removes the build directory entirely then makes fresh checkout from the database.
This remove all other files except those tracked and ignored by Monotone. It will remove all the files that appear in mtn ls unknown. Then it will pull from remote and update the working directory.
This remove all other files except those tracked by Monotone. It will remove all the files that appear in mtn ls ignored and mtn ls unknows. Then pull and update similar to clean
This first checkout source into source directory then copy the source directory to build directory then performs the build operation in the copied directory. This way we make fresh builds with very less bandwidth to download source. The behavior of source checkout follows exactly same as incremental. It performs all the incremental checkout behavior in source directory. ShellCommand

Most interesting steps involve executing a process of some sort on the worker. The ShellCommand class handles this activity.

Several subclasses of ShellCommand are provided as starting points for common build steps.

Using ShellCommands

This is a useful base class for just about everything you might want to do during a build (except for the initial source checkout). It runs a single command in a child shell on the worker. All stdout/stderr is recorded into a LogFile. The step usually finishes with a status of FAILURE if the command's exit code is non-zero, otherwise it has a status of SUCCESS.

The preferred way to specify the command is with a list of argv strings, since this allows for spaces in filenames and avoids doing any fragile shell-escaping. You can also specify the command with a single string, in which case the string is given to /bin/sh -c COMMAND for parsing.

On Windows, commands are run via cmd.exe /c which works well. However, if you're running a batch file, the error level does not get propagated correctly unless you add 'call' before your batch file's name: cmd=['call', 'myfile.bat', ...].

The ShellCommand arguments are:


a list of strings (preferred) or single string (discouraged) which specifies the command to be run. A list of strings is preferred because it can be used directly as an argv array. Using a single string (with embedded spaces) requires the worker to pass the string to /bin/sh for interpretation, which raises all sorts of difficult questions about how to escape or interpret shell metacharacters.

If command contains nested lists (for example, from a properties substitution), then that list will be flattened before it is executed.


All ShellCommands are run by default in the workdir, which defaults to the build subdirectory of the worker builder's base directory. The absolute path of the workdir will thus be the worker's basedir (set as an option to buildbot-worker create-worker, Creating a worker) plus the builder's basedir (set in the builder's builddir key in master.cfg) plus the workdir itself (a class-level attribute of the BuildFactory, defaults to build).

For example:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "test"],

a dictionary of environment strings which will be added to the child command's environment. For example, to run tests with a different i18n language setting, you might use:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "test"],
                             env={'LANG': 'fr_FR'}))

These variable settings will override any existing ones in the worker's environment or the environment specified in the Builder. The exception is PYTHONPATH, which is merged with (actually prepended to) any existing PYTHONPATH setting. The following example will prepend /home/buildbot/lib/python to any existing PYTHONPATH:

from buildbot.plugins import steps

              command=["make", "test"],
              env={'PYTHONPATH': "/home/buildbot/lib/python"}))

To avoid the need of concatenating path together in the master config file, if the value is a list, it will be joined together using the right platform dependant separator.

Those variables support expansion so that if you just want to prepend /home/buildbot/bin to the PATH environment variable, you can do it by putting the value ${PATH} at the end of the value like in the example below. Variables that don't exist on the worker will be replaced by "".

from buildbot.plugins import steps

              command=["make", "test"],
              env={'PATH': ["/home/buildbot/bin",

Note that environment values must be strings (or lists that are turned into strings). In particular, numeric properties such as buildnumber must be substituted using Interpolate.

if False, stdout from the child process is discarded rather than being sent to the buildmaster for inclusion in the step's LogFile.
like want_stdout but for stderr. Note that commands run through a PTY do not have separate stdout/stderr streams: both are merged into stdout.

Should this command be run in a pty? False by default. This option is not available on Windows.

In general, you do not want to use a pseudo-terminal. This is is only useful for running commands that require a terminal - for example, testing a command-line application that will only accept passwords read from a terminal. Using a pseudo-terminal brings lots of compatibility problems, and prevents Buildbot from distinguishing the standard error (red) and standard output (black) streams.

In previous versions, the advantage of using a pseudo-terminal was that grandchild processes were more likely to be cleaned up if the build was interrupted or times out. This occurred because using a pseudo-terminal incidentally puts the command into its own process group.

As of Buildbot-0.8.4, all commands are placed in process groups, and thus grandchild processes will be cleaned up properly.


Sometimes commands will log interesting data to a local file, rather than emitting everything to stdout or stderr. For example, Twisted's trial command (which runs unit tests) only presents summary information to stdout, and puts the rest into a file named _trial_temp/test.log. It is often useful to watch these files as the command runs, rather than using /bin/cat to dump their contents afterwards.

The logfiles= argument allows you to collect data from these secondary logfiles in near-real-time, as the step is running. It accepts a dictionary which maps from a local Log name (which is how the log data is presented in the build results) to either a remote filename (interpreted relative to the build's working directory), or a dictionary of options. Each named file will be polled on a regular basis (every couple of seconds) as the build runs, and any new text will be sent over to the buildmaster.

If you provide a dictionary of options instead of a string, you must specify the filename key. You can optionally provide a follow key which is a boolean controlling whether a logfile is followed or concatenated in its entirety. Following is appropriate for logfiles to which the build step will append, where the pre-existing contents are not interesting. The default value for follow is False, which gives the same behavior as just providing a string filename.

from buildbot.plugins import steps

                   command=["make", "test"],
                   logfiles={"triallog": "_trial_temp/test.log"}))

The above example will add a log named 'triallog' on the master, based on _trial_temp/test.log on the worker.

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "test"],
                                 "triallog": {
                                    "filename": "_trial_temp/test.log",
                                    "follow": True
If set to True, logfiles will be tracked lazily, meaning that they will only be added when and if something is written to them. This can be used to suppress the display of empty or missing log files. The default is False.
if the command fails to produce any output for this many seconds, it is assumed to be locked up and will be killed. This defaults to 1200 seconds. Pass None to disable.
if the command takes longer than this many seconds, it will be killed. This is disabled by default.
If this option is True (the default), then the step's logfile will describe the environment variables on the worker. In situations where the environment is not relevant and is long, it may be easier to set logEnviron=False.
If the command should be interrupted (either by buildmaster or timeout etc.), what signal should be sent to the process, specified by name. By default this is "KILL" (9). Specify "TERM" (15) to give the process a chance to cleanup. This functionality requires a 0.8.6 worker or newer.


If set, when interrupting, try to kill the command with SIGTERM and wait for sigtermTime seconds before firing interuptSignal. If None, interruptSignal will be fired immediately on interrupt.
If the command expects input on stdin, that can be supplied a a string with this parameter. This value should not be excessively large, as it is handled as a single string throughout Buildbot -- for example, do not pass the contents of a tarball with this parameter.
This is a dictionary that decodes exit codes into results value. For example, {0:SUCCESS,1:FAILURE,2:WARNINGS}, will treat the exit code 2 as WARNINGS. The default is to treat just 0 as successful. ({0:SUCCESS}) any exit code not present in the dictionary will be treated as FAILURE
Shell Sequence

Some steps have a specific purpose, but require multiple shell commands to implement them. For example, a build is often configure; make; make install. We have two ways to handle that:

  • Create one shell command with all these. To put the logs of each commands in separate logfiles, we need to re-write the script as configure 1> configure_log; ... and to add these configure_log files as logfiles argument of the buildstep. This has the drawback of complicating the shell script, and making it harder to maintain as the logfile name is put in different places.
  • Create three ShellCommand instances, but this loads the build UI unnecessarily.

ShellSequence is a class to execute not one but a sequence of shell commands during a build. It takes as argument a renderable, or list of commands which are ShellArg objects. Each such object represents a shell invocation.

The single ShellSequence argument aside from the common parameters is:


A list of ShellArg objects or a renderable the returns a list of ShellArg objects.

from buildbot.plugins import steps, util

        util.ShellArg(command=['make'], logfile='make'),
        util.ShellArg(command=['make', 'check_warning'], logfile='warning', warnOnFailure=True),
        util.ShellArg(command=['make', 'install'], logfile='make install')

All these commands share the same configuration of environment, workdir and pty usage that can be setup the same way as in ShellCommand.

class buildbot.steps.shellsequence.ShellArg(self, command=None, logfile=None, haltOnFailure=False, flunkOnWarnings=False, flunkOnFailure=False, warnOnWarnings=False, warnOnFailure=False)
  • command -- (see the ShellCommand command argument),
  • logfile -- optional log file name, used as the stdio log of the command

The haltOnFailure, flunkOnWarnings, flunkOnFailure, warnOnWarnings, warnOnFailure parameters drive the execution of the sequence, the same way steps are scheduled in the build. They have the same default values as for buildsteps - see Common Parameters.

Any of the arguments to this class can be renderable.

Note that if logfile name does not start with the prefix stdio, that prefix will be set like stdio <logfile>.

The two ShellSequence methods below tune the behavior of how the list of shell commands are executed, and can be overridden in subclasses.

class buildbot.steps.shellsequence.ShellSequence
Parameters:oneCmd -- a string or a list of strings, as rendered from a ShellArg instance's command argument.

Determine whether the command oneCmd should be executed. If shouldRunTheCommand returns False, the result of the command will be recorded as SKIPPED. The default methods skips all empty strings and empty lists.


Return the status text of the step in the end. The default value is to set the text describing the execution of the last shell command.

Parameters:commands -- list of shell args

This method actually runs the shell sequence. The default run method calls runShellSequence, but subclasses can override run to perform other operations, if desired.


This is intended to handle the ./configure step from autoconf-style projects, or the perl Makefile.PL step from perl modules. The default command is ./configure but you can change this by providing a command= parameter. The arguments are identical to ShellCommand.

from buildbot.plugins import steps

class buildbot.steps.cmake.CMake

This is intended to handle the cmake step for projects that use CMake-based build systems.


Links below point to the latest CMake documentation. Make sure that you check the documentation for the CMake you use.

In addition to the parameters ShellCommand supports, this step accepts the following parameters:

Either a path to a source directory to (re-)generate a build system for it in the current working directory. Or an existing build directory to re-generate its build system.
A build system generator. See cmake-generators(7) for available options.
A dictionary that contains parameters that will be converted to -D{name}={value} when passed to CMake. Refer to cmake(1) for more information.
A list or a tuple that contains options that will be passed to CMake as is. Refer to cmake(1) for more information.
Path to the CMake binary. Default is cmake
from buildbot.plugins import steps


            'CMAKE_BUILD_TYPE': Property('BUILD_TYPE')


This is meant to handle compiling or building a project written in C. The default command is make all. When the compilation is finished, the log file is scanned for GCC warning messages, a summary log is created with any problems that were seen, and the step is marked as WARNINGS if any were discovered. Through the WarningCountingShellCommand superclass, the number of warnings is stored in a Build Property named warnings-count, which is accumulated over all Compile steps (so if two warnings are found in one step, and three are found in another step, the overall build will have a warnings-count property of 5). Each step can be optionally given a maximum number of warnings via the maxWarnCount parameter. If this limit is exceeded, the step will be marked as a failure.

The default regular expression used to detect a warning is '.*warning[: ].*' , which is fairly liberal and may cause false-positives. To use a different regexp, provide a warningPattern= argument, or use a subclass which sets the warningPattern attribute:

from buildbot.plugins import steps

f.addStep(steps.Compile(command=["make", "test"],
                        warningPattern="^Warning: "))

The warningPattern= can also be a pre-compiled Python regexp object: this makes it possible to add flags like re.I (to use case-insensitive matching).

Note that the compiled warningPattern will have its match method called, which is subtly different from a search. Your regular expression must match the from the beginning of the line. This means that to look for the word "warning" in the middle of a line, you will need to prepend '.*' to your regular expression.

The suppressionFile= argument can be specified as the (relative) path of a file inside the workdir defining warnings to be suppressed from the warning counting and log file. The file will be uploaded to the master from the worker before compiling, and any warning matched by a line in the suppression file will be ignored. This is useful to accept certain warnings (e.g. in some special module of the source tree or in cases where the compiler is being particularly stupid), yet still be able to easily detect and fix the introduction of new warnings.

The file must contain one line per pattern of warnings to ignore. Empty lines and lines beginning with # are ignored. Other lines must consist of a regexp matching the file name, followed by a colon (:), followed by a regexp matching the text of the warning. Optionally this may be followed by another colon and a line number range. For example:

# Sample warning suppression file

mi_packrec.c : .*result of 32-bit shift implicitly converted to 64 bits.* : 560-600
DictTabInfo.cpp : .*invalid access to non-static.*
kernel_types.h : .*only defines private constructors and has no friends.* : 51

If no line number range is specified, the pattern matches the whole file; if only one number is given it matches only on that line.

The default warningPattern regexp only matches the warning text, so line numbers and file names are ignored. To enable line number and file name matching, provide a different regexp and provide a function (callable) as the argument of warningExtractor=. The function is called with three arguments: the BuildStep object, the line in the log file with the warning, and the SRE_Match object of the regexp search for warningPattern. It should return a tuple (filename, linenumber, warning_test). For example:

                  warningPattern="^(.\*?):([0-9]+): [Ww]arning: (.\*)$",

(Compile.warnExtractFromRegexpGroups is a pre-defined function that returns the filename, linenumber, and text from groups (1,2,3) of the regexp match).

In projects with source files in multiple directories, it is possible to get full path names for file names matched in the suppression file, as long as the build command outputs the names of directories as they are entered into and left again. For this, specify regexps for the arguments directoryEnterPattern= and directoryLeavePattern=. The directoryEnterPattern= regexp should return the name of the directory entered into in the first matched group. The defaults, which are suitable for GNU Make, are these:

directoryEnterPattern="make.*: Entering directory [\"`'](.*)['`\"]"
directoryLeavePattern="make.*: Leaving directory"

(TODO: this step needs to be extended to look for GCC error messages as well, and collect them into a separate logfile, along with the source code filenames involved).

Visual C++

These steps are meant to handle compilation using Microsoft compilers. VC++ 6-14 (aka Visual Studio 2003-2015 and VCExpress9) are supported via calling devenv. Msbuild as well as Windows Driver Kit 8 are supported via the MsBuild4, MsBuild12, and MsBuild14 steps. These steps will take care of setting up a clean compilation environment, parsing the generated output in real time, and delivering as detailed as possible information about the compilation executed.

All of the classes are in buildbot.steps.vstudio. The available classes are:

  • VC6
  • VC7
  • VC8
  • VC9
  • VC10
  • VC11
  • VC12
  • VC14
  • VS2003
  • VS2005
  • VS2008
  • VS2010
  • VS2012
  • VS2013
  • VS2015
  • VCExpress9
  • MsBuild4
  • MsBuild12
  • MsBuild14

The available constructor arguments are

The mode default to rebuild, which means that first all the remaining object files will be cleaned by the compiler. The alternate values are build, where only the updated files will be recompiled, and clean, where the current build files are removed and no compilation occurs.
This is a mandatory argument which specifies the project file to be used during the compilation.
This argument defaults to release an gives to the compiler the configuration to use.
This is the place where the compiler is installed. The default value is compiler specific and is the default place where the compiler is installed.
This boolean parameter, defaulting to False instruct the compiler to use its own settings or the one defined through the environment variables PATH, INCLUDE, and LIB. If any of the INCLUDE or LIB parameter is defined, this parameter automatically switches to True.
This is a list of path to be added to the PATH environment variable. The default value is the one defined in the compiler options.
This is a list of path where the compiler will first look for include files. Then comes the default paths defined in the compiler options.
This is a list of path where the compiler will first look for libraries. Then comes the default path defined in the compiler options.
That one is only available with the class VS2005 (VC8). It gives the target architecture of the built artifact. It defaults to x86 and does not apply to MsBuild4 or MsBuild12. Please see platform below.
This gives the specific project to build from within a workspace. It defaults to building all projects. This is useful for building cmake generate projects.
This is a mandatory argument for MsBuild4 and MsBuild12 specifying the target platform such as 'Win32', 'x64' or 'Vista Debug'. The last one is an example of driver targets that appear once Windows Driver Kit 8 is installed.

Here is an example on how to drive compilation with Visual Studio 2013:

from buildbot.plugins import steps

    steps.VS2013(projectfile="project.sln", config="release",
        arch="x64", mode="build",

Here is a similar example using "MsBuild12":

from buildbot.plugins import steps

# Build one project in Release mode for Win32
    steps.MsBuild12(projectfile="trunk.sln", config="Release", platform="Win32",

# Build the entire solution in Debug mode for x64
    steps.MsBuild12(projectfile="trunk.sln", config='Debug', platform='x64',

This step runs cppcheck, analyse its output, and set the outcome in Properties.

from buildbot.plugins import steps

f.addStep(steps.Cppcheck(enable=['all'], inconclusive=True]))

This class adds the following arguments:

(Optional, default to cppcheck) Use this if you need to give the full path to the cppcheck binary or if your binary is called differently.
(Optional, default to ['.']) This is the list of paths for the sources to be checked by this step.
(Optional) Use this to give a list of the message classes that should be in cppcheck report. See the cppcheck man page for more information.
(Optional) Set this to True if you want cppcheck to also report inconclusive results. See the cppcheck man page for more information.
(Optional) This is the list of extra arguments to be given to the cppcheck command.

All other arguments are identical to ShellCommand.

class buildbot.steps.mswin.Robocopy

This step runs robocopy on Windows.

Robocopy is available in versions of Windows starting with Windows Vista and Windows Server 2008. For previous versions of Windows, it's available as part of the Windows Server 2003 Resource Kit Tools.

from buildbot.plugins import steps, util

        description='Deploying binaries...',
        descriptionDone='Deployed binaries.',

Available constructor arguments are:

The path to the source directory (mandatory).
The path to the destination directory (mandatory).
An array of file names or patterns to copy.
Copy files and directories recursively (/E parameter).
Mirror the source directory in the destination directory, including removing files that don't exist anymore (/MIR parameter).
Delete the source directory after the copy is complete (/MOVE parameter).
An array of file names or patterns to exclude from the copy (/XF parameter).
An array of directory names or patterns to exclude from the copy (/XD parameter).
An array of custom parameters to pass directly to the robocopy command.
Whether to output verbose information (/V /TS /TP parameters).

Note that parameters /TEE /NP will always be appended to the command to signify, respectively, to output logging to the console, use Unicode logging, and not print any percentage progress information for each file.

from buildbot.plugins import steps


This is meant to handle unit tests. The default command is make test, and the warnOnFailure flag is set. The other arguments are identical to ShellCommand.

from buildbot.plugins import steps


This is a simple command that uses the du tool to measure the size of the code tree. It puts the size (as a count of 1024-byte blocks, aka 'KiB' or 'kibibytes') on the step's status text, and sets a build property named tree-size-KiB with the same value. All arguments are identical to ShellCommand.

from buildbot.plugins import steps


This is a simple command that knows how to run tests of perl modules. It parses the output to determine the number of tests passed and failed and total number executed, saving the results for later query. The command is prove --lib lib -r t, although this can be overridden with the command argument. All other arguments are identical to those for ShellCommand.

MTR (mysql-test-run)

The MTR class is a subclass of Test. It is used to run test suites using the mysql-test-run program, as used in MySQL, Drizzle, MariaDB, and MySQL storage engine plugins.

The shell command to run the test suite is specified in the same way as for the Test class. The MTR class will parse the output of running the test suite, and use the count of tests executed so far to provide more accurate completion time estimates. Any test failures that occur during the test are summarized on the Waterfall Display.

Server error logs are added as additional log files, useful to debug test failures.

Optionally, data about the test run and any test failures can be inserted into a database for further analysis and report generation. To use this facility, create an instance of twisted.enterprise.adbapi.ConnectionPool with connections to the database. The necessary tables can be created automatically by setting autoCreateTables to True, or manually using the SQL found in the source file.

One problem with specifying a database is that each reload of the configuration will get a new instance of ConnectionPool (even if the connection parameters are the same). To avoid that Buildbot thinks the builder configuration has changed because of this, use the steps.mtrlogobserver.EqConnectionPool subclass of ConnectionPool, which implements an equiality operation that avoids this problem.

Example use:

from buildbot.plugins import steps, util

myPool = util.EqConnectionPool("MySQLdb", "host", "buildbot", "password", "db")
myFactory.addStep(steps.MTR(workdir="mysql-test", dbpool=myPool,
                            command=["perl", "", "--force"]))

The MTR step's arguments are:

Maximum number of test failures to show on the waterfall page (to not flood the page in case of a large number of test failures. Defaults to 5.
Maximum length of test names to show unabbreviated in the waterfall page, to avoid excessive column width. Defaults to 16.
Value of option --parallel option used for (number of processes used to run the test suite in parallel). Defaults to 4. This is used to determine the number of server error log files to download from the worker. Specifying a too high value does not hurt (as nonexisting error logs will be ignored), however if using option --parallel value greater than the default it needs to be specified, or some server error logs will be missing.
An instance of twisted.enterprise.adbapi.ConnectionPool, or None. Defaults to None. If specified, results are inserted into the database using the ConnectionPool.
Boolean, defaults to False. If True (and dbpool is specified), the necessary database tables will be created automatically if they do not exist already. Alternatively, the tables can be created manually from the SQL statements found in the source file.
Short string that will be inserted into the database in the row for the test run. Defaults to the empty string, but can be specified to identify different types of test runs.
Descriptive string that will be inserted into the database in the row for the test run. Defaults to the empty string, but can be specified as a user-readable description of this particular test run.
The subdirectory in which to look for server error log files. Defaults to mysql-test, which is usually correct. Interpolate is supported.
class buildbot.steps.subunit.SubunitShellCommand

This buildstep is similar to ShellCommand, except that it runs the log content through a subunit filter to extract test and failure counts.

from buildbot.plugins import steps

f.addStep(steps.SubunitShellCommand(command="make test"))

This runs make test and filters it through subunit. The 'tests' and 'test failed' progress metrics will now accumulate test data from the test run.

If failureOnNoTests is True, this step will fail if no test is run. By default failureOnNoTests is False. Worker Filesystem Steps

Here are some buildsteps for manipulating the worker's filesystem.


This step will assert that a given file exists, failing if it does not. The filename can be specified with a property.

from buildbot.plugins import steps


This step requires worker version 0.8.4 or later.


This command copies a directory on the worker.

from buildbot.plugins import steps

f.addStep(steps.CopyDirectory(src="build/data", dest="tmp/data"))

This step requires worker version 0.8.5 or later.

The CopyDirectory step takes the following arguments:

if the copy command fails to produce any output for this many seconds, it is assumed to be locked up and will be killed. This defaults to 120 seconds. Pass None to disable.
if the command takes longer than this many seconds, it will be killed. This is disabled by default.

This command recursively deletes a directory on the worker.

from buildbot.plugins import steps


This step requires worker version 0.8.4 or later.


This command creates a directory on the worker.

from buildbot.plugins import steps


This step requires worker version 0.8.5 or later. Python BuildSteps

Here are some BuildSteps that are specifically useful for projects implemented in Python.

class buildbot.steps.python.BuildEPYDoc

epydoc is a tool for generating API documentation for Python modules from their docstrings. It reads all the .py files from your source tree, processes the docstrings therein, and creates a large tree of .html files (or a single .pdf file).

The BuildEPYDoc step will run epydoc to produce this API documentation, and will count the errors and warnings from its output.

You must supply the command line to be used. The default is make epydocs, which assumes that your project has a Makefile with an epydocs target. You might wish to use something like epydoc -o apiref source/PKGNAME instead. You might also want to add option --pdf to generate a PDF file instead of a large tree of HTML files.

The API docs are generated in-place in the build tree (under the workdir, in the subdirectory controlled by the option -o argument). To make them useful, you will probably have to copy them to somewhere they can be read. For example if you have server with configured nginx web server, you can place generated docs to it's public folder with command like rsync -ad apiref/ You might instead want to bundle them into a tarball and publish it in the same place where the generated install tarball is placed.

from buildbot.plugins import steps

f.addStep(steps.BuildEPYDoc(command=["epydoc", "-o", "apiref", "source/mypkg"]))
class buildbot.steps.python.PyFlakes

PyFlakes is a tool to perform basic static analysis of Python code to look for simple errors, like missing imports and references of undefined names. It is like a fast and simple form of the C lint program. Other tools (like pychecker) provide more detailed results but take longer to run.

The PyFlakes step will run pyflakes and count the various kinds of errors and warnings it detects.

You must supply the command line to be used. The default is make pyflakes, which assumes you have a top-level Makefile with a pyflakes target. You might want to use something like pyflakes . or pyflakes src.

from buildbot.plugins import steps

f.addStep(steps.PyFlakes(command=["pyflakes", "src"]))
class buildbot.steps.python.Sphinx

Sphinx is the Python Documentation Generator. It uses RestructuredText as input format.

The Sphinx step will run sphinx-build or any other program specified in its sphinx argument and count the various warnings and error it detects.

from buildbot.plugins import steps


This step takes the following arguments:

(required) Name of the directory where the documentation will be generated.
(optional, defaulting to .), Name the directory where the file will be found
(optional) Indicates the builder to use.
(optional, defaulting to sphinx-build) Indicates the executable to run.
(optional) List of tags to pass to sphinx-build
(optional) Dictionary of defines to overwrite values of the file.
(optional) String, one of full or incremental (the default). If set to full, indicates to Sphinx to rebuild everything without re-using the previous build results.

Similarly, the PyLint step will run pylint and analyze the results.

You must supply the command line to be used. There is no default.

from buildbot.plugins import steps

f.addStep(steps.PyLint(command=["pylint", "src"]))
class buildbot.steps.python_twisted.Trial

This step runs a unit test suite using trial, a unittest-like testing framework that is a component of Twisted Python. Trial is used to implement Twisted's own unit tests, and is the unittest-framework of choice for many projects that use Twisted internally.

Projects that use trial typically have all their test cases in a 'test' subdirectory of their top-level library directory. For example, for a package petmail, the tests might be in petmail/test/test_*.py. More complicated packages (like Twisted itself) may have multiple test directories, like twisted/test/test_*.py for the core functionality and twisted/mail/test/test_*.py for the email-specific tests.

To run trial tests manually, you run the trial executable and tell it where the test cases are located. The most common way of doing this is with a module name. For petmail, this might look like trial petmail.test, which would locate all the test_*.py files under petmail/test/, running every test case it could find in them. Unlike the that comes with Python, it is not necessary to run the as a script; you always let trial do the importing and running. The step's tests` parameter controls which tests trial will run: it can be a string or a list of strings.

To find the test cases, the Python search path must allow something like import petmail.test to work. For packages that don't use a separate top-level lib directory, PYTHONPATH=. will work, and will use the test cases (and the code they are testing) in-place. PYTHONPATH=build/lib or PYTHONPATH=build/lib.somearch are also useful when you do a python build step first. The testpath attribute of this class controls what PYTHONPATH is set to before running trial.

Trial has the ability, through the --testmodule flag, to run only the set of test cases named by special test-case-name tags in source files. We can get the list of changed source files from our parent Build and provide them to trial, thus running the minimal set of test cases needed to cover the Changes. This is useful for quick builds, especially in trees with a lot of test cases. The testChanges parameter controls this feature: if set, it will override tests.

The trial executable itself is typically just trial, and is typically found in the shell search path. It can be overridden with the trial parameter. This is useful for Twisted's own unittests, which want to use the copy of bin/trial that comes with the sources.

To influence the version of Python being used for the tests, or to add flags to the command, set the python parameter. This can be a string (like python2.2) or a list (like ['python2.3', '-Wall']).

Trial creates and switches into a directory named _trial_temp/ before running the tests, and sends the twisted log (which includes all exceptions) to a file named test.log. This file will be pulled up to the master where it can be seen as part of the status output.

from buildbot.plugins import steps


Trial has the ability to run tests on several workers in parallel (beginning with Twisted 12.3.0). Set jobs to the number of workers you want to run. Note that running trial in this way will create multiple log files (named test.N.log, err.N.log and out.N.log starting with N=0) rather than a single test.log.

This step takes the following arguments:

(optional) Number of worker-resident trial workers to use when running the tests. Defaults to 1 worker. Only works with Twisted>=12.3.0.
class buildbot.steps.python_twisted.RemovePYCs

This is a simple built-in step that will remove .pyc files from the workdir. This is useful in builds that update their source (and thus do not automatically delete .pyc files) but where some part of the build process is dynamically searching for Python modules. Notably, trial has a bad habit of finding old test modules.

from buildbot.plugins import steps

f.addStep(steps.RemovePYCs()) Transferring Files
class buildbot.steps.transfer.FileUpload
class buildbot.steps.transfer.FileDownload

Most of the work involved in a build will take place on the worker. But occasionally it is useful to do some work on the buildmaster side. The most basic way to involve the buildmaster is simply to move a file from the worker to the master, or vice versa. There are a pair of steps named FileUpload and FileDownload to provide this functionality. FileUpload moves a file up to the master, while FileDownload moves a file down from the master.

As an example, let's assume that there is a step which produces an HTML file within the source tree that contains some sort of generated project documentation. And let's assume that we run nginx web server on buildmaster host for serving static files. We want to move this file to the buildmaster, into a /usr/share/nginx/www/ directory, so it can be visible to developers. This file will wind up in the worker-side working directory under the name docs/reference.html. We want to put it into the master-side /usr/share/nginx/www/ref.html, and add a link to the HTML status to the uploaded file.

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "docs"]))

The masterdest= argument will be passed to os.path.expanduser, so things like ~ will be expanded properly. Non-absolute paths will be interpreted relative to the buildmaster's base directory. Likewise, the workersrc= argument will be expanded and interpreted relative to the builder's working directory.


The copied file will have the same permissions on the master as on the worker, look at the mode= parameter to set it differently.

To move a file from the master to the worker, use the FileDownload command. For example, let's assume that some step requires a configuration file that, for whatever reason, could not be recorded in the source code repository or generated on the worker side:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "config"]))

Like FileUpload, the mastersrc= argument is interpreted relative to the buildmaster's base directory, and the workerdest= argument is relative to the builder's working directory. If the worker is running in ~worker, and the builder's builddir is something like tests-i386, then the workdir is going to be ~worker/tests-i386/build, and a workerdest= of foo/bar.html will get put in ~worker/tests-i386/build/foo/bar.html. Both of these commands will create any missing intervening directories.

Other Parameters

The maxsize= argument lets you set a maximum size for the file to be transferred. This may help to avoid surprises: transferring a 100MB coredump when you were expecting to move a 10kB status file might take an awfully long time. The blocksize= argument controls how the file is sent over the network: larger blocksizes are slightly more efficient but also consume more memory on each end, and there is a hard-coded limit of about 640kB.

The mode= argument allows you to control the access permissions of the target file, traditionally expressed as an octal integer. The most common value is probably 0755, which sets the x executable bit on the file (useful for shell scripts and the like). The default value for mode= is None, which means the permission bits will default to whatever the umask of the writing process is. The default umask tends to be fairly restrictive, but at least on the worker you can make it less restrictive with a --umask command-line option at creation time (Worker Options).

The keepstamp= argument is a boolean that, when True, forces the modified and accessed time of the destination file to match the times of the source file. When False (the default), the modified and accessed times of the destination file are set to the current time on the buildmaster.

The url= argument allows you to specify an url that will be displayed in the HTML status. The title of the url will be the name of the item transferred (directory for DirectoryUpload or file for FileUpload). This allows the user to add a link to the uploaded item if that one is uploaded to an accessible place.

Transfering Directories
class buildbot.steps.transfer.DirectoryUpload

To transfer complete directories from the worker to the master, there is a BuildStep named DirectoryUpload. It works like FileUpload, just for directories. However it does not support the maxsize, blocksize and mode arguments. As an example, let's assume an generated project documentation, which consists of many files (like the output of doxygen or epydoc). And let's assume that we run nginx web server on buildmaster host for serving static files. We want to move the entire documentation to the buildmaster, into a /usr/share/nginx/www/docs directory, and add a link to the uploaded documentation on the HTML status page. On the worker-side the directory can be found under docs:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "docs"]))

The DirectoryUpload step will create all necessary directories and transfers empty directories, too.

The maxsize and blocksize parameters are the same as for FileUpload, although note that the size of the transferred data is implementation-dependent, and probably much larger than you expect due to the encoding used (currently tar).

The optional compress argument can be given as 'gz' or 'bz2' to compress the datastream.


The permissions on the copied files will be the same on the master as originally on the worker, see option buildbot-worker create-worker --umask to change the default one.

Transferring Multiple Files At Once
class buildbot.steps.transfer.MultipleFileUpload

In addition to the FileUpload and DirectoryUpload steps there is the MultipleFileUpload step for uploading a bunch of files (and directories) in a single BuildStep. The step supports all arguments that are supported by FileUpload and DirectoryUpload, but instead of a the single workersrc parameter it takes a (plural) workersrcs parameter. This parameter should either be a list, or something that can be rendered as a list.:

from buildbot.plugins import steps

f.addStep(steps.ShellCommand(command=["make", "test"]))
f.addStep(steps.ShellCommand(command=["make", "docs"]))
f.addStep(steps.MultipleFileUpload(workersrcs=["docs", "test-results.html"],

The url= parameter, can be used to specify a link to be displayed in the HTML status of the step.

The way URLs are added to the step can be customized by extending the MultipleFileUpload class. The allUploadsDone method is called after all files have been uploaded and sets the URL. The uploadDone method is called once for each uploaded file and can be used to create file-specific links.

import os

from buildbot.plugins import steps

class CustomFileUpload(steps.MultipleFileUpload):
    linkTypes = ('.html', '.txt')

    def linkFile(self, basename):
        name, ext = os.path.splitext(basename)
        return ext in self.linkTypes

    def uploadDone(self, result, source, masterdest):
        if self.url:
            basename = os.path.basename(source)
            if self.linkFile(basename):
                self.addURL(self.url + '/' + basename, basename)

    def allUploadsDone(self, result, sources, masterdest):
        if self.url:
            notLinked = filter(lambda src: not self.linkFile(src), sources)
            numFiles = len(notLinked)
            if numFiles:
                self.addURL(self.url, '... %d more' % numFiles) Transfering Strings
class buildbot.steps.transfer.StringDownload
class buildbot.steps.transfer.JSONStringDownload
class buildbot.steps.transfer.JSONPropertiesDownload

Sometimes it is useful to transfer a calculated value from the master to the worker. Instead of having to create a temporary file and then use FileDownload, you can use one of the string download steps.

from buildbot.plugins import steps, util


StringDownload works just like FileDownload except it takes a single argument, s, representing the string to download instead of a mastersrc argument.

from buildbot.plugins import steps

buildinfo = {
    'branch': Property('branch'),
    'got_revision': Property('got_revision')
f.addStep(steps.JSONStringDownload(buildinfo, workerdest="buildinfo.json"))

JSONStringDownload is similar, except it takes an o argument, which must be JSON serializable, and transfers that as a JSON-encoded string to the worker.

from buildbot.plugins import steps


JSONPropertiesDownload transfers a json-encoded string that represents a dictionary where properties maps to a dictionary of build property name to property value; and sourcestamp represents the build's sourcestamp. Running Commands on the Master
class buildbot.steps.master.MasterShellCommand

Occasionally, it is useful to execute some task on the master, for example to create a directory, deploy a build result, or trigger some other centralized processing. This is possible, in a limited fashion, with the MasterShellCommand step.

This step operates similarly to a regular ShellCommand, but executes on the master, instead of the worker. To be clear, the enclosing Build object must still have a worker object, just as for any other step -- only, in this step, the worker does not do anything.

In this example, the step renames a tarball based on the day of the week.

from buildbot.plugins import steps

    command="mv widgetsoft-new.tar.gz widgetsoft-`date +%a`.tar.gz",


By default, this step passes a copy of the buildmaster's environment variables to the subprocess. To pass an explicit environment instead, add an env={..} argument.

Environment variables constructed using the env argument support expansion so that if you just want to prepend /home/buildbot/bin to the PATH environment variable, you can do it by putting the value ${PATH} at the end of the value like in the example below. Variables that don't exist on the master will be replaced by "".

from buildbot.plugins import steps

              command=["make", "www"],
              env={'PATH': ["/home/buildbot/bin",

Note that environment values must be strings (or lists that are turned into strings). In particular, numeric properties such as buildnumber must be substituted using Interpolate.

(optional) The directory from which the command will be ran.
(optional) Signal to use to end the process, if the step is interrupted.
class buildbot.steps.master.LogRenderable

This build step takes content which can be renderable and logs it in a pretty-printed format. It can be useful for debugging properties during a build. Setting Properties

These steps set properties on the master based on information from the worker.

class buildbot.steps.master.SetProperty

SetProperty takes two arguments of property and value where the value is to be assigned to the property key. It is usually called with the value argument being specifed as a Interpolate object which allows the value to be built from other property values:

from buildbot.plugins import steps, util

        value=util.Interpolate("sch=%(prop:scheduler)s, worker=%(prop:workername)s")

This buildstep is similar to ShellCommand, except that it captures the output of the command into a property. It is usually used like this:

from buildbot.plugins import steps

f.addStep(steps.SetPropertyFromCommand(command="uname -a", property="uname"))

This runs uname -a and captures its stdout, stripped of leading and trailing whitespace, in the property uname. To avoid stripping, add strip=False.

The property argument can be specified as a Interpolate object, allowing the property name to be built from other property values.

Passing includeStdout=False (default True) stops capture from stdout.

Passing includeStderr=True (default False) allows capture from stderr.

The more advanced usage allows you to specify a function to extract properties from the command output. Here you can use regular expressions, string interpolation, or whatever you would like. In this form, extract_fn should be passed, and not Property. The extract_fn function is called with three arguments: the exit status of the command, its standard output as a string, and its standard error as a string. It should return a dictionary containing all new properties.

Note that passing in extract_fn will set includeStderr to True.

def glob2list(rc, stdout, stderr):
    jpgs = [l.strip() for l in stdout.split('\n')]
    return {'jpgs': jpgs}

f.addStep(SetPropertyFromCommand(command="ls -1 *.jpg",