-
Bug
-
Resolution: Won't Do
-
Normal
-
None
-
4.16
-
None
-
False
-
Description of problem:
A cluster that has MachineSets that don't have their own MachineConfigPools (e.g. 'infra' and 'workload') is converted to TechPreview. After TechPreview is rolled out, co/machine-config emits an error about "MachineConfigNode ... is invalid: spec.pool.name: Invalid value" If I manually create the relevant MachineConfigNode and MachineConfigPools objects after the changes roll out, the CO error does not disappear. The errors persist after machine-config-controller and machine-config-operator pod restarts.
Version-Release number of selected component (if applicable):
4.16.0-0.nightly-2024-07-17-092440
How reproducible:
100% (today & yesterday. Two new clusters)
Steps to Reproduce:
1. Install 4.16 cluster 2. Create new MachineSets 3. Convert to TechPreviewNoUpgrade 4. See errors in co/machine-config
Actual results:
See errors in co/machine-config. Creating the objects does not remove the errors.
Error looks like:
# oc get co machine-config
machine-config 4.16.0-0.nightly-2024-07-17-092440 True False True 101m Failed to resync 4.16.0-0.nightly-2024-07-17-092440 because: MachineConfigNode.machineconfiguration.openshift.io "ip-10-0-31-80.us-west-2.compute.internal" is invalid: spec.pool.name: Invalid value: "": spec.pool.name in body should match '^([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9])(\.([a-zA-Z0-9]|[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}[a-zA-Z0-9]))*$'
Expected results:
We may see the errors after TechPreview rolls out, but the controller should pick up the MCN and MCP after they are created and reconcile.
Additional info:
I have a m-g available
