[Feature] Safe-guard Multi-node etcd cluster members against wrongly mounted volumes
ishan16696 opened this issue · comments
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 write POD_NAME
in safe_guard
file.
/assign