-
Notifications
You must be signed in to change notification settings - Fork 38.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Various fixes to job #14142
Various fixes to job #14142
Conversation
42962f4
to
1225756
Compare
@k8s-bot test this please |
GCE e2e build/test passed for commit 12257569221a2c79d5860333c8f776a5b4de63e5. |
Labelling this PR as size/S |
GCE e2e build/test passed for commit 12257569221a2c79d5860333c8f776a5b4de63e5. |
Labelling this PR as size/L |
GCE e2e build/test passed for commit 70baa11c0f0d5493347b7a0a8b4e929545a23bf3. |
LGTM |
needs rebase |
3af20e4
to
e29e606
Compare
GCE e2e build/test passed for commit e29e606. |
@@ -403,10 +398,6 @@ func (jm *JobManager) manageJob(activePods []*api.Pod, successful, unsuccessful | |||
} else if active < parallelism { | |||
// how many executions are left to run | |||
diff := *job.Spec.Completions - successful | |||
// for RestartPolicyNever we need to count unsuccessful pods as well |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why this was removed? Without it, the comment in types.go does not make sense, since we'll always update the unsuccessful value. Besides, this PR also change the idea of not restarting failed pods when RestartPolicyNever is set. Currently the RestartPolicy does not change behavior of job controller, is that indented? May I know the reasoning?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How job interacts with restart policy was not part of the initial proposal. We really wanted to get job in but we wanted to discuss this interaction in design review over the next couple months while it was in experimental. Thanks for pointing out that incorrect comment, but with the logic the way it was this comment is incorrect because both successful pods and failed pod will be counted towards my completion.
Please see #14186 which is an umbrella issue for jobcontroller enhancements.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks, I'll move entire discussion there. Besides thanks for all the fixes you've made in the mean time to the jobs!
No description provided.