-
Bug
-
Resolution: Done
-
Major
-
None
-
4.12.z
-
None
-
No
-
SDN Sprint 234, SDN Sprint 235, SDN Sprint 236
-
3
-
Rejected
-
False
-
In the node healthz server, leveraged by cloud load balancers to check the node health for services with ExternalTrafficPolicy=Cluster, we should respond negatively to health checks as soon as the ovnk node pod is to be deleted, by inspecting the pod DeletionTimestamp.
This allows cloud load balancers to have a little more time to update their list of healthy nodes before the node actually stops altogether forwarding traffic.
- depends on
-
OCPBUGS-10913 Node healthz server: return unhealthy when pod is to be deleted
- Closed
- is depended on by
-
OCPBUGS-10915 Node healthz server: return unhealthy when pod is to be deleted
- Closed
- links to