-
Bug
-
Resolution: Won't Do
-
Major
-
None
-
4.14, 4.15, 4.16
-
Quality / Stability / Reliability
-
False
-
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
The OpenShift Node Performance Operator leverages MachineConfigs to apply kernelArgs and systemd services.
The MachineConfigOperator only supports managing boot-configuration and kernel arguments for RHCOS Nodes.
https://github.com/openshift/machine-config-operator/blob/master/docs/FAQ.md#q-does-the-mco-run-on-rhel-worker-nodes
Additionally, another limitation is called out in the ClusterNodeTuningOperator Github documentation but is not referenced or discussed in our customer-facing documentation.
https://github.com/openshift/cluster-node-tuning-operator?tab=readme-ov-file#additional-tuning-on-fully-managed-hosts
We should note that some of the tuning features are not supported/working for RHEL Nodes and list which features are not tested.
Version-Release number of selected component (if applicable):
4.16
How reproducible:
Everytime
Steps to Reproduce:
1. Create an OpenShift Cluster
2. Add RHEL Nodes to the OpenShift cluster and create an MCP for them
3. Configure a PerformanceProfile for the RHEL-Node MCP
4. The MachineConfig is generated and rolled-out
Actual results:
The MachineConfig is rolled-out but the kernel arguments are never seen in the Node.
Expected results:
It still fails but it's noted as failing in the OpenShift Documentation.
Additional info:
