Cluster

A storage cluster is operated through its StorageCluster resource. A cluster-wide action is requested by setting spec.action, and the Simplyblock Operator calls the control plane and reports the outcome in the status of the resource. The topology settings of a cluster, its failure domains and its node affinity, are properties of the same resource and are read when data placement is decided.

Topic	Purpose
Storage Cluster Actions	The mechanism behind every action: requesting, executing, re-running, monitoring.
Activating a Storage Cluster	Making a cluster serve I/O when the automatic activation did not happen.
Shutting Down a Storage Cluster	Suspending the whole cluster and taking all of its volumes offline.
Starting a Storage Cluster	Bringing a suspended cluster back into service.
Restarting a Storage Cluster	Sequencing a shutdown and a start in one action.
Rolling Restart	Restarting every storage node of a cluster, one node at a time.
Upgrading a Cluster	Upgrading the Helm chart, and rolling the storage plane node by node afterward.
Managing Failure Domains	Grouping storage nodes into fault groups, so that chunks are spread across them.
Configuring Node Affinity	Keeping the data of a logical volume on the storage node that owns it.