Fix ImagePullTimeout to use "Initialized" #7882
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #7881.
ImagePullBackOffTimeOut currently uses the "Scheduled" pod condition transition time, however a pod can be scheduled on a node far earlier than when image pulling begins which can (and does) result in Tasks failing when image pulls could be retried and succeed.
We should use the transition time for "PodReadyToStartContainers" (1.29+) / "Initialized" for checking the image pull timeout as these conditions more accurately capture when the image pull began.
Changes
This commit updates the imagePullBackOffTimeOut logic to use the "PodInitialized" or "PodReadyToStartContainers" Pod Condition in favour of "PodScheduled" and adds the covering tests.
Submitter Checklist
As the author of this PR, please check off the items in this checklist:
/kind <type>. Valid types are bug, cleanup, design, documentation, feature, flake, misc, question, tepRelease Notes