Details
-
Bug
-
Resolution: Can't Do
-
Major
-
None
-
4.7
-
None
-
Important
-
No
-
Rejected
-
False
-
Description
Description of problem:
Customer is using a Jenkins application that connects to the externally deployed OCS. When the connection to External server is lost the containers aren't able to use the existing PVCs to claim the volume. Manual restart of affected pods is required to make things normal
Version-Release number of selected component (if applicable):
OCP v4.7
How reproducible:
Easily reproducible
Steps to Reproduce:
1. Use any externally deployed storage(OCS, NFS etc) 2. Scale up pods & bind storage using PVCs 3. Restart the external storage to interrupt the connection. 4. Now, the liveliness probe will fail & container will be restarted but not the pod. 5. Pod remains stuck untill re-started to establish the connection again.
Actual results:
Pod in CrashLoopBackOff as container failing health check probes
Expected results:
Pod restarting as a whole not just the container
Additional info:
I had suggested them to point liveliness & readiness probe to same URL with hope that it will re-start the pod & not just the container. But, it didn't helped.