OpenShift Container Platform

Evicting KubeVirt virtual machines

In cases where KubeVirt virtual machines (VMs) cannot be live migrated, such as when you use GPU passthrough, the VMs must be evicted at the same time as the NodePool resource of the hosted cluster.

Otherwise, the compute nodes might be shut down without being drained from the workload. This might also happen when you are upgrading the OpenShift Virtualization Operator.

To achieve a synchronized restart, you can set the evictionStrategy parameter on the hyperconverged resource to ensure that only VMs that are drained from workloads are rebooted.

Procedure

To learn more about the hyperconverged resource and the allowed values for the evictionStrategy parameter, enter the following command:
```
$ oc explain hyperconverged.spec.virtualization.evictionStrategy
```

Patch the hyperconverged resource by entering the following command:

$ oc patch -n openshift-cnv hco kubevirt-hyperconverged \
  --type=merge \
  -p '{"spec": {"virtualization": {"evictionStrategy": "External"}}}'

Patch the workload update strategy and the workload update methods by entering the following command:
```
$ oc patch -n openshift-cnv hco kubevirt-hyperconverged \
  --type=merge \
  -p '{"spec": {"virtualization": {"workloadUpdateStrategy": {"workloadUpdateMethods": ["LiveMigrate","Evict"]}}}}'
```
By applying this patch, you specify that VMs should be live-migrated if possible, and that only the VMs that cannot be live-migrated should be evicted.

Verification

Check whether the patch command was applied properly by entering the following command:

$ oc get -n openshift-cnv hco kubevirt-hyperconverged -ojsonpath='{.spec.virtualization.evictionStrategy}'

Example output

External