A job configuration using sjb will take one of three forms: a stand-alone job,
one that inherits from another or one that triggers other jobs. A standalone
job can declare options; a job with a parent can extend or override any option
from the parent.
Parent-child relationships can nest -- to generate a job, the topmost parent configuration is loaded, then iteratively extended and overridden by children configurations until the bottommost configuration has been evaluated.
description is an optional string field that will override the default job
description. The syntax is:
description: ""children is a list which, if present, will configure the job to have child
jobs in a multi-job build. If this field is present, the only other acceptable
configuration fields are timer, email, and merge or test. The syntax
is:
children: [] # a list of job names (YAML filenames without extension under sjb/config/test_cases)parent is an optional string field that takes already existing
job configuration and includes it into the current job configuration.
The syntax is:
parent: "" # path to a job configuration with respect to sjb/config directoryIt is used together with extensions and overrides keyword that specifies
a list of configuration options (see example at the bottom).
A configuration option can either be found in extensions: or in
overrides but not in both.
extensions is a dictionary used together with parent field that extends
a dictionary of parent's configuration options.
The syntax is:
parent: ""
extensions: {} # a dictionary of configuration optionsoverrides is a dictionary used together with parent field that replaces
a dictionary of parent's configuration options.
The syntax is:
parent: ""
overrides: {} # a dictionary of configuration optionstimer is an optional field that holds a cron entry and will configure the
job to trigger on this timer. Normal Jenkins cron syntax is expected. The
syntax is:
timer: "" # cron entry, e.g. 'H H * * *'email is an optional list of e-mail addresses for who to contact when the job
fails. The syntax is:
email: [] # list of e-mail addressesjunit_analysis is an optional flag to enable or disable the "Publish
JUnit test result report" post-build step. It defaults to True
junit_analysis: False # JUnit analysis is disabledmerge and test are optional fields that mark the job as one that uses the
test-pull-requests utility for signalling that a test succeeded or merging
a pull request given a test success. If test-pull-requests is configured to
run a job for a [test] tag, the test field should be present in the job
configuration. Likewise with [merge] and the merge field. The syntax is:
merge: "" # repo name under github.com/openshifttest: "" # repo name under github.com/openshiftNote: one of merge or test can be specified, but not both.
parameters is an optional field which adds parameters to the Jenkins job. The
field contains a list of parameter definitions. The parameter definition syntax
is:
parameters:
- name: "" # environment variable name
description: "" # human-readable description
default_value: "" # default value [optional]provision is a required field which gives the options for provisioning a VM
in AWS EC2 for the job. The syntax is:
provision:
os: "" # operating system. only 'rhel' is supported
stage: "" # image stage. one of ['bare', 'base', 'build', 'install', 'fork']
provider: "" # cloud provider. only 'aws' is supportedsync_repos is an optional field which declares which repositories are to be
synced and how they are to be synced. This field will add build stages to sync
the repository as well as update the job description with details about what
repos the job instance is being run with. The syntax is:
sync_repos:
- name: "" # repository name under github.com/openshift
type: "" # 'pull_request' for a PR sync [optional]actions is an optional list of actions to take in the job. Each action will
declare a type and can optionally declare a timeout. The syntax is:
actions:
- type: "" # one of ['forward_parameters', 'host_script', 'script']
timeout: 0 # in seconds [optional]the forward_parameters action will make an environment variable available on
the Jenkins master in the $WORKSPACE also available to scripts running on the
remote host in EC2. The syntax is:
actions:
- type: "forward_parameters"
timeout: 0 # in seconds [optional]
parameters: [] # a list of environment variablesthe host_script action will run a set of shell commands on the Jenkins master
in the job's $WORKSPACE. The syntax is:
actions:
- type: "host_script"
timeout: 0 # in seconds [optional]
title: "" # a human-readable title for the step
script: "" # inline shell script to run, no character escaping necessarythe script action will run a set of shell commands in the specified directory
on the remote VM in EC2. The syntax is:
actions:
- type: "script"
timeout: 0 # in seconds [optional]
title: "" # a human-readable title for the step
repository: "" # the repository under github.com/openshift to `cd` into [optional]
script: "" # inline shell script to run, no character escaping necessarypost_actions is an optional list of actions to be taken after the main job
actions have finished or errored. The syntax is identical to that for actions:
post_actions:
- type: "" # one of ['forward_parameters', 'host_script', 'script']artifacts is an optional list of files or directories on the remote VM which
will be retrieved into the $WORKSPACE on the Jenkins master after the job is
finished. The syntax is:
artifacts: [] # a list of absolute pathsgenerated_artifacts is an optional mapping of shell commands to artifact file
names. For each mapping entry, the command will be run on the remote VM and the
output captured to the filename in the $WORKSPACE on the Jenkins master. The
syntax is:
generated_artifacts:
filename: "shell command" # note: this is a map, not a listsystem_journals is an optional list of systemd units for which the system
journal should be gathered from the remote VM. Each unit name will be used to
expand: journalctl --unit {{ unit }} --no-pager --all --lines=all. The syntax
is:
system_journals: [] # a list of unit names (e.g. 'docker')This job definition serves as a proto-job, used as a parent for all other
Origin test jobs. It lives in config/common.
parameters: []
provision:
os: "rhel"
stage: "base"
provider: "aws"
sync_repos:
- name: "origin"
actions:
- type: "script"
title: "use a ramdisk for etcd"
timeout: 300
script: |-
sudo su root <<SUDO
mkdir -p /tmp
mount -t tmpfs -o size=4096m tmpfs /tmp
mkdir -p /tmp/etcd
chmod a+rwx /tmp/etcd
restorecon -R /tmp
echo "ETCD_DATA_DIR=/tmp/etcd" >> /etc/environment
SUDO
post_actions: []
artifacts:
- "/data/src/github.com/openshift/origin/_output/scripts"
generated_artifacts:
installed_packages.log: 'sudo yum list installed'
avc_denials.log: 'sudo ausearch -m AVC -m SELINUX_ERR -m USER_AVC'
docker.info: 'sudo docker version && sudo docker info && sudo docker images && sudo docker ps -a'
filesystem.info: 'sudo df -h && sudo vgs && sudo lvs'
pid1.journal: 'sudo journalctl _PID=1 --no-pager --all --lines=all'
system_journals:
- docker.service
This job definition defines the job that runs unit tests and verification for
Origin. It inherits from the above origin.yml job and lives in config.
parent: 'common/test_cases/origin.yml'
extensions:
actions:
- type: "script"
title: "verify commit history"
repository: "origin"
timeout: 7200
script: |-
# run commitchecker outside release container as it needs
# access to git; also explicitly force godeps verification
branch="$( git rev-parse --abbrev-ref --symbolic-full-name HEAD )"
if [[ "${branch}" == "master" ]]; then
RESTORE_AND_VERIFY_GODEPS=1 make verify-commits -j
fi
- type: "script"
title: "run check and verify tasks"
repository: "origin"
script: |-
OS_BUILD_ENV_PRESERVE=_output/scripts hack/env TEST_KUBE='true' JUNIT_REPORT='true' make check -j -kNote: a configuration option can either be found in extensions: or in
overrides but not in both.