-
Spike
-
Resolution: Done
-
Critical
-
None
-
None
-
None
-
Incidents & Support
-
False
-
-
False
-
None
-
None
-
None
-
OCP Console - Sprint 277
Which 4.y.z to 4.y'.z' updates increase vulnerability?
Updates into 4.19.(7 <= z <= 13 ), until OCPBUGS-61879 ships a 4.19.z fix.
Which types of clusters?
Clusters are only exposed if the configured plugin set:
$ oc get -o jsonpath-as-json='{.spec.plugins[*]}' console.operator.openshift.io cluster
includes a name that lacks a ConsolePlugin:
$ oc get -o name consoleplugins.console.openshift.io
If a ConsolePlugin exists for every spec.plugins entry, the cluster is not exposed.
What is the impact? Is it serious enough to warrant removing update recommendations?
This bug can cause the console Pods to crash-loop, blocking update completion. But the PodDisruptionBudget on the console Pods will keep the cluster from removing the final functioning console Pod until the cluster admins recover the cluster.
How involved is remediation?
Remove the broken plugin from the cluster console.operator.openshift.io's spec.plugins before starting the update to avoid running into the issue:
$ oc patch console.operator.openshift.io cluster --type json -p '[{"op": "remove", "path": "/spec/plugins/FIXME"}]'
where you should replace FIXME with the index of the plugin you want to remove (where 0 references the first plugin).
The same operation will also resolve the issue if the cluster already run into it during or after update.
Alternatively, update your cluster to a version that includes the OCPBUGS-61879 fix.
Is this a regression?
Yes, this was introduced in 4.19.7, where clusters with spec.plugin references that lacked an associated ConsolePlugin resource were not tested while verifying a fix for another issue.
- blocks
-
OCPBUGS-60782 Enabling invalid or missing plugin causes console to crash loop
-
- Verified
-
- links to