If you are using a released version of Kubernetes, you should refer to the docs that go with that version.
The latest release of this document can be found [here](http://releases.k8s.io/release-1.3/docs/devel/automation.md).Documentation for other releases can be found at releases.k8s.io.
Kubernetes uses a variety of automated tools in an attempt to relieve developers of repetitive, low brain power work. This document attempts to describe these processes.
In an effort to
- reduce load on core developers
- maintain e2e stability
- load test githubs label feature
We have added an automated [submit-queue] (https://github.com/kubernetes/contrib/blob/master/mungegithub/mungers/submit-queue.go) to the github "munger" for kubernetes.
The submit-queue does the following:
for _, pr := range readyToMergePRs() {
if testsAreStable() {
if retestPR(pr) == success {
mergePR(pr)
}
}
}
The status of the submit-queue is online.
The submit-queue lists what it believes are required on the merge requirements tab of the info page. That may be more up to date.
A PR is considered "ready for merging" if it matches the following:
- The PR must have the label "cla: yes" or "cla: human-approved"
- The PR must be mergeable. aka cannot need a rebase
- All of the following github statuses must be green
- Jenkins GCE Node e2e
- Jenkins GCE e2e
- Jenkins unit/integration
- The PR cannot have any prohibited future milestones (such as a v1.5 milestone during v1.4 code freeze)
- The PR must have the "lgtm" label
- The PR must not have been updated since the "lgtm" label was applied
- The PR must not have the "do-not-merge" label
Merges only occur when the critical builds are passing. We're open to including more builds here, let us know...
Merges are serialized, so only a single PR is merged at a time, to ensure against races.
If the PR has the retest-not-required
label, it is simply merged. If the PR does
not have this label the e2e, unit/integration, and node tests are re-run. If these
tests pass a second time, the PR will be merged as long as the critical builds
are
green when this PR finishes retesting.
We run a github "munger".
This runs repeatedly over github pulls and issues and runs modular "mungers" similar to "mungedocs." The mungers include the 'submit-queue' referenced above along with numerous other functions. See the README in the link above.
Please feel free to unleash your creativity on this tool, send us new mungers that you think will help support the Kubernetes development process.
We also run a robotic PR builder that attempts to run tests for each PR.
Before a PR from an unknown user is run, the PR builder bot (k8s-bot
) asks to
a message from a contributor that a PR is "ok to test", the contributor replies
with that message. ("please" is optional, but remember to treat your robots with
kindness...)
PRs should only need to be manually re-tested if you believe there was a flake during the original test. All flakes should be filed as an issue. Once you find or file a flake a contributer (this may be you!) should request a retest with "@k8s-bot test this issue: #NNNNN", where NNNNN is replaced with the issue number you found or filed.
Any pushes of new code to the PR will automatically trigger a new test. No human interraction is required.