Major Description of problem:
We, on openshift-virtualization, are running this command:
oc adm must-gather --dest-dir=./tests-collected-info/must_gather_scope_module/collected_vm_gather_from_vm_node --image=registry.redhat.io/container-native-virtualization/cnv-must-gather-rhel9@sha256:13425c8ab8b4c584c8666d938ab677dbf1a584aad1e0c4a9cb7c44c027a8cefa --node-name=c01-hmeir421-fn6lh-worker-0-nrnpc -- /usr/bin/gather --vms_details
and the pod is never scheduled because the nodeName and the affinity definitions on the must-gather pod spec are mutually exclusive.
Version-Release number of selected component (if applicable):
$ oc version Client Version: 4.21.0-ec.2 Kustomize Version: v5.6.0 Server Version: 4.21.0-ec.2 Kubernetes Version: v1.34.1
How reproducible:
Always
Actual results:
the must-gather pod is never scheduled
Expected results:
the must-gather pod is run on the node selected through --node-name
Additional info:
must-gather pod details attached
- is triggered by
-
CNV-70614 [stabilization]Must-gather on specific worker killed in non-compact clusters
-
- New
-
- links to