-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
4.13.z
-
Moderate
-
No
-
False
-
-
-
-
30-April; cu's are closing support cases as soon a workaround is applied; workaround is removing gatewayConfig from CO before patching. KCS Published
-
-
-
Description of problem:
The upgrade to z-stream version of 4.13 is failing due to the error. (in case of the error below from 4.13.10 to 4.13.37)
///
Error while trying to update operator configuration: failed to apply / update (operator.openshift.io/v1, Kind=Network) /cluster: Network.operator.openshift.io "cluster" is invalid: spec.defaultNetwork.ovnKubernetesConfig.gatewayConfig: Invalid value: "null": spec.defaultNetwork.ovnKubernetesConfig.gatewayConfig in body must be of type object: "null"
The upgrade would not progress and the CNO logs show same thing.
///
2024-04-05T21:53:45.245016281Z I0405 21:53:45.244976 1 log.go:198] Could not propagate configuration from network.config.openshift.io to network.operator.openshift.io: failed to apply / update (operator.openshift.io/v1, Kind=Network) /cluster: Network.operator.openshift.io "cluster" is invalid: spec.defaultNetwork.ovnKubernetesConfig.gatewayConfig: Invalid value: "null": spec.defaultNetwork.ovnKubernetesConfig.gatewayConfig in body must be of type object: "null"
We fixed the issue by running a workaround to remove the configuration.
# oc patch network.operator.openshift.io cluster --type=json -p="[{'op': 'remove', 'path': '/spec/defaultNetwork/ovnKubernetesConfig/gatewayConfig'}]"
However, the we need to know the cause of that issue.
Why would the CNO be stuck and why the original patch with error "Invalid value: "null" was presented.
CNO should reconcile the latest version and should progress.
Similar issue was shown before, however with different error.
///
Invalid value: "null": spec.defaultNetwork.ovnKubernetesConfig.hybridOverlayConfig.hybridClusterNetwork in body must be of type array: "null"
Same workaround was applied.
Version-Release number of selected component (if applicable):
OpenShift Container Platform 4.13.10 -> 4.13.37
How reproducible:
N/A at this moment, we are not able to reproduce the issue.
Steps to Reproduce:
1.
2.
3.
Actual results:
Upgrade is stuck on CNO.
Expected results:
The CNO should reconcile and not show that error. The cause of the error is unknown. The upgrade should processed normally.
Additional info:
The workaround fixed the issue and upgrade finished just fine.
The issue was only visible on their prod system - not on their preprod. Hence the request of RCA.
