don't do pod deletion management for pipeline builds #10370
Conversation
|
@csrwng ptal |
|
@csrwng bump |
|
@bparees sorry I missed this one earlier. What does it mean when your pipeline build is still running but a pod associated with the build is deleted? |
| @@ -373,9 +373,13 @@ func (bc *BuildPodDeleteController) HandleBuildPodDeletion(pod *kapi.Pod) error | |||
| } | |||
| build := obj.(*buildapi.Build) | |||
|
|
|||
| if build.Spec.Strategy.JenkinsPipelineStrategy != nil { | |||
There was a problem hiding this comment.
you and your high standards. yeah let me do that.
it means craziness happened because there should never be a pod associated w/ a pipeline build. but because of some of the weird things we do to handle resync scenarios, it's possible a "pod deletion" event could get generated for a pod that never existed, which could cause us to mark the pipeline build as "ERROR/FAIL" (which is what we normally do when a pod disappears for a running build). so i'm just trying to make very sure we don't do anything with pipeline builds as part of the pod deletion handling logic. (and conversely, if the pipeline build is deleted, that we don't accidentally terminate an innocent pod) |
|
got it |
bparees
force-pushed
the
pipeline_build_deletion
branch
from
30fa932 to
1ad5d0f
Compare
bparees
force-pushed
the
pipeline_build_deletion
branch
from
1ad5d0f to
fb6397a
Compare
|
@csrwng UTed. |
|
LGTM |
|
[merge] Ben Parees | OpenShift On Aug 12, 2016 5:25 PM, "Cesar Wong" [email protected] wrote:
|
|
continuous-integration/openshift-jenkins/merge SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pr_origin/7897/) (Image: devenv-rhel7_4821) |
|
Evaluated for origin merge up to fb6397a |
No description provided.