[Feature] Safe-guard Multi-node etcd cluster members against wrongly mounted volumes #566
Labels
kind/enhancement
Enhancement, improvement, extension
lifecycle/rotten
Nobody worked on this for 12 months (final aging stage)
priority/2
Priority (lower number equals higher priority)
Feature (What you would like to be added):
Right now, I have observed that backup-restore is only writing POD Namespace in safe_guard file as shoot identifier to avoid the issue of wrong volume mount which I feel is not enough in case of multi-node etcd cluster(size>1) as if volume of one etcd cluster member can wrongly mounted to other etcd member of same etcd cluster.
Right now, backup-restore won't able to recognise this case as
POD_NAMESPACE
of both etcd cluster member is same, so the content of safe_guard file will also be same.Expected behavior:
etcd backup-restore should not start etcd cluster member if volume of any other etcd cluster member get wrongly mounted.
Approach/Hint to the implement solution (optional):
Along with
POD_NAMESPACE
, we can also writePOD_NAME
insafe_guard
file.The text was updated successfully, but these errors were encountered: