Details
-
Bug
-
Resolution: Won't Do
-
Major
-
None
-
4.14
-
No
-
SDN Sprint 240, SDN Sprint 241, SDN Sprint 242, SDN Sprint 243, SDN Sprint 244, SDN Sprint 245, SDN Sprint 246, SDN Sprint 247, SDN Sprint 248
-
9
-
Rejected
-
False
-
Description
Description of problem:
Using telemetry data and running an intense workload on a 51 node ROSA cluster, we observed that the peak memory usage in some containers of the ovnkube-node pod is less than the requests, so it perhaps makes sense to adjust the requests to create additional capacity for the workload pods on workers.
Version-Release number of selected component (if applicable):
4.14.0-0.nightly-2023-08-08-094653
How reproducible:
100%
Steps to Reproduce:
1. Use kube-burner to run this workload on a ROSA/Self-managed OCP cluster 2. https://github.com/smalleni/kube-burner/commit/d2b4f20f1de20ca0e70d71070331ae61e15698a0 3. kube-burner ocp cluster-density-v2 --iterations=650
You can also reproduce this by running the regular clcuster-density-v2 workload with churn=true
Actual results:
Max container memory in MiB during testing: northd: 85 ovn-controller-95 nbdb: 40 sbdb: 55
Expected results:
Additional info:
