Keep nodes while deorbiter is running #720
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
deleting volumes in the deorbiter is failing for multiple minutes because the nodes are gone and the csi daemonset is not unmounting the volumes. This changes lets the launch controller wait for the deorbiter to finish work before starting to terminate nodes. For the volumes to actually be deleted we need to remove pods that hold a reference to the pvc.
databus23
requested review from
BugRoger,
onuryilmaz and
jknipper
as code owners
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Deleting volumes in the deorbiter is failing for multiple minutes because the nodes are gone and the csi daemonset is not unmounting the volumes.
This leads to very long cluster deletion times in our soak test:
This changes lets the launch controller wait for the deorbiter to finish work before starting to terminate nodes.
For the volumes to actually be deleted we need to remove pods that hold a reference to the pvc how that the nodes stay around.
With this change the deletion time is down consideribly:
(I think deletion can be even faster, its currently limited by the 120s wait for loadbalancer deletion)
Open question: