While investigating a problem with OpenShift Container Platform 4 - Node scaling, I found the below messages reported in my OpenShift Container Platform 4 - Cluster.

E0513 11:15:09.331353       1 orchestrator.go:450] Couldn't get autoscaling options for ng: MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c
E0513 11:15:09.331365       1 orchestrator.go:450] Couldn't get autoscaling options for ng: MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.331529       1 orchestrator.go:546] Pod project-100/curl-67f84bd857-h92wb can't be scheduled on MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c, predicate checking error: node(s) didn't match Pod's node affinity/selector; predicateName=NodeAffinity; reasons: node(s) didn't match Pod's node affinity/selector; debugInfo=
I0513 11:15:09.331684       1 orchestrator.go:157] No pod can fit to MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c
E0513 11:15:09.332076       1 orchestrator.go:507] Failed to get autoscaling options for node group MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c: Not implemented
I0513 11:15:09.332100       1 orchestrator.go:185] Best option to resize: MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332110       1 orchestrator.go:189] Estimated 1 nodes needed in MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332135       1 orchestrator.go:295] Final scale-up plan: [{MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c 0->1 (max: 12)}]

The same events are reported in must-gather reviewed from customers. Given that we have https://github.com/kubernetes/autoscaler/issues/6037 and https://github.com/kubernetes/autoscaler/issues/6676 that appear to be solved via https://github.com/kubernetes/autoscaler/pull/6677 and https://github.com/kubernetes/autoscaler/pull/6038 I'm wondering whether we should pull in those changes as they seem to eventually impact automated scaling of OpenShift Container Platform 4 - Node(s).

OpenShift Container Platform 4.15

Always

1. Setup OpenShift Container Platform 4 with ClusterAutoscaler configured
2. Trigger scaling activity and verify the cluster-autoscaler-default logs

Logs like the below are being reported.

E0513 11:15:09.331353       1 orchestrator.go:450] Couldn't get autoscaling options for ng: MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c
E0513 11:15:09.331365       1 orchestrator.go:450] Couldn't get autoscaling options for ng: MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.331529       1 orchestrator.go:546] Pod project-100/curl-67f84bd857-h92wb can't be scheduled on MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c, predicate checking error: node(s) didn't match Pod's node affinity/selector; predicateName=NodeAffinity; reasons: node(s) didn't match Pod's node affinity/selector; debugInfo=
I0513 11:15:09.331684       1 orchestrator.go:157] No pod can fit to MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c
E0513 11:15:09.332076       1 orchestrator.go:507] Failed to get autoscaling options for node group MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c: Not implemented
I0513 11:15:09.332100       1 orchestrator.go:185] Best option to resize: MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332110       1 orchestrator.go:189] Estimated 1 nodes needed in MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332135       1 orchestrator.go:295] Final scale-up plan: [{MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c 0->1 (max: 12)}]

Scale-up of OpenShift Container Platform 4 - Node to happen without error being reported

I0513 11:15:09.331529       1 orchestrator.go:546] Pod project-100/curl-67f84bd857-h92wb can't be scheduled on MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c, predicate checking error: node(s) didn't match Pod's node affinity/selector; predicateName=NodeAffinity; reasons: node(s) didn't match Pod's node affinity/selector; debugInfo=
I0513 11:15:09.331684       1 orchestrator.go:157] No pod can fit to MachineSet/openshift-machine-api/test-12345-batch-amd64-us-east-2c
I0513 11:15:09.332100       1 orchestrator.go:185] Best option to resize: MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332110       1 orchestrator.go:189] Estimated 1 nodes needed in MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c
I0513 11:15:09.332135       1 orchestrator.go:295] Final scale-up plan: [{MachineSet/openshift-machine-api/test-12345-batch-arm64-us-east-2c 0->1 (max: 12)}]

Please review https://github.com/kubernetes/autoscaler/issues/6037 and https://github.com/kubernetes/autoscaler/issues/6676 as they seem to document the problem and also have a solution linked/merged