Description of problem:
When trying to destroy the OCP cluster on GCP, I hit a Panic error. The error is a nil pointer dereference in the waitFor method of the Cluster Uninstaller struct. The panic occurs in the waitFor method, which is called by handleOperation, invoked by deleteDisk, and ultimately part of destroyDisks. The likely cause is a nil pointer being passed to waitFor, possibly due to a failed GCP API call or an uninitialized client.
Version-Release number of selected component (if applicable):
4.18
How reproducible:
Steps to Reproduce:
1. Install 4.18 cluster on GCP 2. Destroy the cluster using openshift-installer
Actual results:
Cluster and associated resources should destroy without any error
Expected results:
During the deletion of cluster resources, hitting a panic error
Additional info:
This seems to occur when a call to the Delete function returns no error but a nil operation (op) like here https://github.com/t-cas/openshift-installer/blob/main/pkg/destroy/gcp/disk.go#L126C32-L126C38 for disks or here https://github.com/t-cas/openshift-installer/blob/main/pkg/destroy/gcp/instance.go#L108 for VMs. Error is triggered when calling this waitFor function https://github.com/openshift/installer/blob/main/pkg/destroy/gcp/gcp.go#L425 as "op" parameter is dereferenced without any check here https://github.com/openshift/installer/blob/main/pkg/destroy/gcp/gcp.go#L442
- blocks
-
OCPBUGS-62981 Panic error during the destruction of an OpenShift 4.18 on GCP using the OpenShift installer
-
- POST
-
- is cloned by
-
OCPBUGS-62981 Panic error during the destruction of an OpenShift 4.18 on GCP using the OpenShift installer
-
- POST
-
- links to