Details
-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.16.0
-
Moderate
-
No
-
False
-
Description
Description of problem:
Version-Release number of selected component (if applicable):
4.16
How reproducible:
During upgrade, it happens frequently
Steps to Reproduce:
1. Launch upgrade 2. Observer monitor test results specicially looking at test outcome of "[sig-node] pods should never transition back to pending"
Actual results:
Flakes for a number of pods that should have an exception
Expected results:
Ignores set of pods that should be allowed to go from Running to Pending
Additional info:
There was a bug that described this problem: https://bugzilla.redhat.com/show_bug.cgi?id=1933760 Outcome was https://github.com/kubernetes/kubernetes/issues/106390 Test should not flake for pods that meet the criteria defined in the above issue. I created this issue because of: https://redhat-internal.slack.com/archives/CE4L0F143/p1704459757540179 David eads: " TODO here is to have the monitor allow the transition for daemonset pods with init containers shortly after a node restart."