Restarting a Storage Cluster

A restart sequences a shutdown and a start of the whole storage cluster, since the control plane offers no cluster restart of its own. Both halves are driven by one StorageClusterOps, and the operation succeeds once the cluster is active again.

Example of a cluster restart (restart-cluster.yaml)

apiVersion: storage.simplyblock.io/v1alpha2
kind: StorageClusterOps
metadata:
  name: restart-simplyblock-cluster
  namespace: simplyblock
spec:
  clusterRef: simplyblock-cluster
  action: Restart

Restarting the storage cluster

kubectl apply -f restart-cluster.yaml

The half that is currently running is the step of the operation.

Step	Description
ShuttingDown	The shutdown is requested, and the step holds until the cluster is no longer active.
Starting	The start is requested, and the step holds until the cluster is active again.

Following the step of a running restart

kubectl get storageclusterops restart-simplyblock-cluster -n simplyblock \
    -o jsonpath='{.status.step.state}{" "}{.status.message}{"\n"}'

Neither step can be aborted, because the cluster is already going down or coming back. The operation cannot be deleted until it is terminal.

Warning

Every volume of the cluster is offline between the shutdown and the start. To restart the storage nodes one after another instead, so that the cluster keeps serving I/O, see Rolling Restart.

How an operation is tracked and cleaned up is described in Storage Cluster Actions.