I did not manage to find any elegant solutions for this.
I solved it with a bash script in a cronjob that runs every 10 minutes.
It checks if any of the nodes are "NotReady" and just force deletes all the pods with "Terminating" status on those nodes, it also deletes all of the volumeattachments assigned to that node.
q
quick-river-12881
06/12/2024, 3:26 PM
You might want to check the Longhorn setting for
Pod Deletion Policy When Node is Down
.
🙌 1
s
shy-shoe-31898
06/13/2024, 10:05 AM
I have this set to:
shy-shoe-31898
06/13/2024, 10:07 AM
but that that does not help as the pods still get stuck at terminating