-
Bug
-
Resolution: Unresolved
-
Critical
-
4.19.z, 4.20.z
-
Quality / Stability / Reliability
-
False
-
-
3
-
None
-
None
-
None
-
None
-
MCO Sprint 277
-
1
-
In Progress
-
Release Note Not Required
-
None
-
None
-
None
-
None
-
None
The MCN CRD cleanup has a couple of issues that ROSA classic folks ran into:
- It uses the newer node-role.kubernetes.io/control-plane: "" in its nodeselector. On clusters that have old master nodes, this could pose an issue, since they'll likely only have the legacy node-role.kubernetes.io/master: "" label. This will prevent the cronjob from getting scheduled.
- It only fails the job when the delete operation fails. For the commands that actually check for the problematic CRD, we currently eat the error. We should log the error in these cases so we're able to debug the root cause.
---------------------------------------
As a workaround to get the upgrade moving , go ahead and add the label as below
$ oc label node <control-plane-node-name> node-role.kubernetes.io/control-plane=
- depends on
-
OCPBUGS-62082 Dummy 4.21 bug for fixing MCN CRD cronjob issue
-
- Verified
-
- is blocked by
-
MCO-1890 Impact statement request for OCPBUGS-62073 MCN CRD clean up script should use legacy master node label and log all errors
-
- Closed
-
- links to