-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.12.z
-
No
-
False
-
Description of problem:
mhc controller seems weirdly busy, always trying to find targets to reconcile, while there is nothing to reconcile. All machines are up and mcp properly applied.
Version-Release number of selected component (if applicable):
4.12.36
How reproducible:
Unsure, we're seeing that on 300+ nodes cluster on openstack. Here 1 mhc selects 33 machines
Steps to Reproduce:
1. 2. 3.
Actual results:
Unsure, I would have expected 1 check for each mhc every 5 minute. For this particular mhc we're seeing one log: machinehealthcheck_controller.go:192] Reconciling openshift-machine-api/xxxx-cdbfg-worker-b2-xxxx: finding targets every 2 seconds. Overall we're seeing 280 "finding targets" per minute, and 3000 "machinehealthcheck_controller.go:469] Reconciling openshift-machine-api/xxxx/yyyy/zzzz: health checking per minute
Expected results:
reasonably less checks per minute
Additional info:
I believe it could be facilitate race conditions as mentioned in OCPBUGS-33584
- links to