jenkins/hudson job configuration generator
Simplify configuration process for multiple projects by building jenkins/hudson job configuration from a template.
Say you have 10 projects of similar type. Be that ant, maven, gradle, grails or projects of any other type. They usually have almost identical configuration: log rotation, source control management, triggers, goals (ant build, maven install, gradle build, grails war, ...), reports, publishers, etc.
Example, let one maven project be configured as follows:
- keep last 10 builds
- use git as SCM
- poll SCM every 10 minutes
- trigger build at midnight every day
- default maven goal is
install
- use reporters: DRY, PMD, FindBugs, CheckStyle with configured thresholds
- mail to
developers@acme.com
on build failure - use cobertura publisher
- collect
target/*.jar
artifacts
Ideally, all other maven projects should reuse this configuration while changing:
- SCM URL
- group id
- artifact id
Manually synchronizing configuration across all projects is a waste of
time. kin
allows you to create templates that are used across projects while
specifying only those properties that differ.
# show help
$ java -jar kin.jar --help
# show version
$ java -jar kin.jar --version
# build job configurations
$ java -jar kin.jar
There is no installation procedure. Simply download latest
distribution, extract zip, create
build.kin
file and run java -jar kin.jar
.
Just Java 6. Everything else is bundled in jar itself.
In order to simplify configuration process kin
uses the fact that
jenkins/hudson store their job configuration in XML file. So when you create new
job named foo
via web interface jenkins stores job configuration in
$JENKINS_HOME/jobs/foo/config.xml
file. $JENKINS_HOME
location can vary.
Jenkins installed by DEB package on Ubuntu uses /var/lib/jenkins
. Running
jenkins standalone uses $HOME/.jenkins
. More information on jenkins wiki
page. The
same rule applies for hudson.
To create jenkins/hudson job configuration via kin
you need to write
build.kin
file and create necessary templates. Usually one per project type
(ant, maven, gradle, grails, etc.) though that is completely up to you.
Easiest way to create a template is to use jenkins/hudson web interface. Create
a new job and configure it the way you'd like. Afterwards, copy created job
configuration into *.tpl
file. For example, for maven project named foo
copy
$JENKINS_HOME/jobs/foo/config.xml
into maven.tpl
. Now edit template by
replacing variable configuration properties with $name
. It will make more
sense in walkthrough section.
Simple DSL is setup to support build.kin
file. It looks like this:
foo { // <- project name
template = 'maven.tpl'
mavenVersion = '3.0.3'
scm = 'http://acme.com/git/foo'
groupId = 'com.acme.foo'
artifactId = 'foo'
}
To share common properties across job configurations:
maven {
// doesn't produce jenkins/hudson job configuration but is only used
// to share common configuration across jobs
producesConfig = false
template = 'maven.tpl' // this is the default
mavenVersion = '3.0.3'
}
foo {
inherit 'maven'
scm = 'http://acme.com/git/foo'
groupId = 'com.acme.foo'
artifactId = 'foo'
}
bar {
inherit 'maven'
scm = 'http://acme.com/git/bar'
groupId = 'com.acme.bar'
artifactId = 'bar'
}
Underneath that DSL is groovy. Therefor we could write more groovy like code to reduce number of lines further:
maven {
producesConfig = false
mavenVersion = '3.0.3'
}
def mavenProject = { name ->
return {
inherit 'maven'
scm = "http://acme.com/git/$name"
groupId = "com.acme.$name"
artifactId = name
}
}
foo mavenProject('foo')
bar mavenProject('bar')
Or you can use addJob
method as part of API that
hr.helix.kin.script.BuildScript
provides:
maven {
producesConfig = false
mavenVersion = '3.0.3'
}
def mavenProject = { name ->
addJob name, {
inherit 'maven'
scm = "http://acme.com/git/$name"
groupId = "com.acme.$name"
artifactId = name
}
}
mavenProject 'foo'
mavenProject 'bar'
You can step away from DSL if you want to. We've already seen addJob
method
available in hr.helix.kin.script.BuildScript
. This block:
foo {
a = 1
b = 2
}
is equivalent to:
addJob 'foo', {
a = 1
b = 2
}
or:
addJob('foo') {
a = 1
b = 2
}
hr.helix.kin.model.Job
class provides inherit
method. It can be used to
specified job name from which to inherit parameters. In following example foo
job inherits both mavenVersion
and groupId
properties.
maven {
producesConfig = false
mavenVersion = '3.0.3'
groupId = 'com.acme'
}
foo {
inherit 'maven'
artifactId = 'foo-all'
}
You can also inherit from multiple jobs. Example:
// omitted maven and git configurations
foo {
inherit 'maven', 'git'
}
More API information can be found by looking at API documentation and source
code (it's really small) available at docs/api
and src
directories
respectively. Following classes are directly responsible for build.kin
file:
hr.helix.kin.script.BuildScript
hr.helix.kin.model.Job
hr.helix.kin.model.Build
Suppose a build.kin
file:
gradle {
producesConfig = false
gradleVersion = '1.0-milestone-4'
task = 'build'
}
kin {
inherit 'gradle'
scm = 'https://github.com/mbezjak/kin'
}
foo {
inherit 'gradle'
scm = 'http://acme.com/git/foo'
task = 'jar'
}
bar {
inherit 'gradle'
gradleVersion = '0.9.2'
scm = 'http://acme.com/git/bar'
}
gradle.tpl
looks like:
<?xml version='1.0' encoding='UTF-8'?>
<project>
<actions/>
<description></description>
<keepDependencies>false</keepDependencies>
<properties/>
<scm class="hudson.plugins.git.GitSCM">
<configVersion>2</configVersion>
<userRemoteConfigs>
<hudson.plugins.git.UserRemoteConfig>
<name></name>
<refspec></refspec>
<url>$scm</url>
</hudson.plugins.git.UserRemoteConfig>
</userRemoteConfigs>
<branches>
<hudson.plugins.git.BranchSpec>
<name>**</name>
</hudson.plugins.git.BranchSpec>
</branches>
<recursiveSubmodules>false</recursiveSubmodules>
<doGenerateSubmoduleConfigurations>false</doGenerateSubmoduleConfigurations>
<authorOrCommitter>false</authorOrCommitter>
<clean>false</clean>
<wipeOutWorkspace>false</wipeOutWorkspace>
<pruneBranches>false</pruneBranches>
<remotePoll>false</remotePoll>
<buildChooser class="hudson.plugins.git.util.DefaultBuildChooser"/>
<gitTool>Default</gitTool>
<submoduleCfg class="list"/>
<relativeTargetDir></relativeTargetDir>
<excludedRegions></excludedRegions>
<excludedUsers></excludedUsers>
<gitConfigName></gitConfigName>
<gitConfigEmail></gitConfigEmail>
<skipTag>false</skipTag>
<scmName></scmName>
</scm>
<canRoam>true</canRoam>
<disabled>false</disabled>
<blockBuildWhenDownstreamBuilding>false</blockBuildWhenDownstreamBuilding>
<blockBuildWhenUpstreamBuilding>false</blockBuildWhenUpstreamBuilding>
<triggers class="vector">
<hudson.triggers.TimerTrigger>
<spec>*/10 * * * *</spec>
</hudson.triggers.TimerTrigger>
</triggers>
<concurrentBuild>false</concurrentBuild>
<builders>
<hudson.plugins.gradle.Gradle>
<description></description>
<switches></switches>
<tasks>build</tasks>
<rootBuildScriptDir></rootBuildScriptDir>
<buildFile></buildFile>
<gradleName>$gradleVersion</gradleName>
<useWrapper>false</useWrapper>
</hudson.plugins.gradle.Gradle>
</builders>
<publishers/>
<buildWrappers/>
</project>
After running java -jar kin.jar
three job configurations (kin, foo, bar) are
created: build/kin/config.xml
, build/foo/config.xml
and
build/bar/config.xml
. When building job configuration, each $name
property
in gradle.tpl
is replaced with actual property value specified in build.kin
.
kin
job configuration:
$gradleVersion
is replaced with1.0-milestone-4
$scm
is replaced withhttps://github.com/mbezjak/kin
$task
is replaced withbuild
foo
job configuration:
$gradleVersion
is replaced with1.0-milestone-4
$scm
is replaced withhttp://acme.com/git/foo
$task
is replaced withjar
bar
job configuration:
$gradleVersion
is replaced with0.9.2
$scm
is replaced withhttp://acme.com/git/bar
$task
is replaced withbuild
Note: every template can also use special jobName
property. It is always bound
to current job configuration. For example, when building job configuration
foo
, property $jobName
has a value of foo
. Same for bar
and so on.
What to do after kin
successfully creates jenkins/hudson job configurations?
Upload created build
directory to $JENKINS_HOME/jobs
directory, of
course. build
directory matches directory
structure of
$JENKINS_HOME/jobs
.
Here is a very simple and elegant solution that works when jenkins/hudson is
installed on Linux. Moreover it automates building configuration files. Put
kin.jar
, build.kin
and *.tpl
files into one directory under SCM (git
,
mercurial
or any other). Create new jenkins/hudson job, setup SCM and build
triggers (example, poll every 10 minutes). SCM should use clean build
option. Add new shell build step with following content:
java -jar "$WORKSPACE/kin.jar"
rsync --archive --checksum --verbose "$WORKSPACE/build/" "$HOME/jobs"
curl "${JENKINS_URL}reload"
Now every time you commit (and push) a change in configuration jenkins/hudson will update configuration and reload itself automatically. Note that this approach hasn't been tested in distributed environment!
Hints for other solutions: rsync, scp, wget, nfs, windows (samba) share, jenkins/hudson script console, jenkins/hudson remote API, etc.
For hudson just replace JENKINS
with HUDSON
.
Remember that build.kin
is powered by very simple DSL. Underneath all that is
still groovy. Unlike *.properties
or *.ini
files when writing build.kin
you have whole programming language to work with.
Templates are processed by
SimpleTemplateEngine
- so groovy again. This gives you a great deal of flexibility when creating templates and building job configurations.
Source code is available at github: https://github.com/mbezjak/kin
Roadmap and changelog are available at: https://github.com/mbezjak/kin/issues
Additional documentation is available at: https://github.com/mbezjak/kin/wiki
Project uses MIT license. Check LICENSE file for more info.