Details
-
Bug
-
Resolution: Fixed
-
Test Blocker
-
None
-
2.6.0
-
None
-
Pre-Christmas Completions, Christmas and Beyond, 1 - 2024 A New Dawn
-
0
Description
We have a couchbase cluster with 3 replicas using CNG. In each pod, CNG fails for different reasons according to the logs.
Log summary:
- In pod 0000 we see protostellar starting and then receiving SIGTERM for failing liveness probe.
- In pod 0001 CNG fails with `failed to reconfigure cbauth` then receiving SIGTERM for failing liveness probe (HTTP probe failed with statuscode: 500)
- In pod 0002 CNG starts and then receiving SIGTERM for failing liveness probe (
HTTP probe failed with statuscode: 404)
I have attached logs for all three containers as they display different issues during startup before being terminated. I also attached operator logs, cluster describe result and a truncated part of k8s events.
Looking at the operator logs, I see a lot of `unknown field \"spec.servers[0].pod.spec.containers\"`, but last time I checked this is a message should not result in failure.
General details:
EKS - 1.25
Server: 7.2.4-7045
Operator: 2.6.0-158
CNG: 0.2.0-142
Probably one of the issue is related to K8S-3288