Details
-
Bug
-
Status: Open
-
Medium
-
Resolution: Unresolved
-
1.14.0
-
None
-
None
-
Yes
-
Yes
Description
There's this ticket for PBM: https://jira.percona.com/browse/PBM-979
After it gets fixed we need to do our fixes for these types of nodes and then re-test.
If we have at least one non-voting member in replica set it will be impossible to run restore - it will be stuck in requested state forever.
$ k get psmdb
NAME ENDPOINT STATUS AGE
my-cluster-name my-cluster-name-rs0.test.svc.cluster.local initializing 26m
$ k get pods
NAME READY STATUS RESTARTS AGE
my-cluster-name-rs0-0 1/1 Running 0 12m
my-cluster-name-rs0-1 1/1 Running 0 12m
my-cluster-name-rs0-2 1/1 Running 0 13m
my-cluster-name-rs0-nv-0 2/2 Running 0 26m
percona-server-mongodb-operator-84675fc76f-fb6jb 1/1 Running 0 26m
$ k get psmdb-restore
NAME CLUSTER STATUS AGE
restore1 my-cluster-name requested 14m
Normal replica set nodes were restarted with new sts, but non voting one was not (as you can see the "nv" pod still has two containers).