-
Bug
-
Resolution: Duplicate
-
Undefined
-
None
-
4.12, 4.14
-
No
-
False
-
-
Release Note Not Required
-
In Progress
This is a clone of issue OCPBUGS-31461. The following is the description of the original issue:
—
This is a clone of issue OCPBUGS-25744. The following is the description of the original issue:
—
Description of problem:
Deleting the node with the Ingress VIP using oc delete node causes a keepalived split-brain
Version-Release number of selected component (if applicable):
4.12, 4.14
How reproducible:
100%
Steps to Reproduce:
1. In an OpenShift cluster installed via vSphere IPI, check the node with the Ingress VIP. 2. Delete the node. 3. Check the discrepancy between machines objects and nodes. There will be more machines than nodes. 4. SSH to the deleted node, and check the VIP is still mounted and keepalived pods are running. 5. Check the VIP is also mounted in another worker. 6. SSH to the node and check the VIP is still present.
Actual results:
The deleted node still has the VIP present and the ingress fails sometimes
Expected results:
The deleted node should not have the VIP present and the ingress should not fail.
Additional info:
- clones
-
-
- Closed
-
- is blocked by
-
-
- Closed
-
