Closed szvincze closed 2 years ago
Here is the proposed fix for the issue.
I managed to reproduce the situation by the method mentioned in the above comment and now the log contains the proper info and there is no pod restart:
INFO controller.nsm Failed to update status {"reconciler group": "nsm.networkservicemesh.io", "reconciler kind": "NSM", "name": "nsm-sample-registry-k8s", "namespace": "nsm", "nsm": "nsm/nsm-sample-registry-k8s", "Error": "Operation cannot be fulfilled on nsms.nsm.networkservicemesh.io \"nsm-sample-registry-k8s\": the object has been modified; please apply your changes to the latest version and try again"}
Describe the bug Sometimes nsm-operator pod restarts when deploying the
config/samples/nsm_v1alpha1_nsm-registry-k8s.yaml
example. The cause of the issue is a panic because of a segmentation violation (please see the attached log for more details):To Reproduce I observed the fault during development of an integration test case, but it was quite rare. However it can be reproduced by deploying and deleting the above mentioned example by the following commands several times in a row:
Expected behavior No pod restart.
Screenshots, CLI output and logs nsm-operator-pod-restart.log
Desktop (please complete the following information):
Additional context