Hardware components (nodes and disks) can be removed from a cluster or reconfigured in other ways when conditions warrant it.

Note: If the Data-at-Rest Encryption is enabled then before removing a drive or node from a cluster, test the certificates again by clicking Test all nodes and ensure that testing is successful and the status is Verified. In case of an SED drive or node, if the drive or node is not removed as recommended then the drive or node will be locked.

1. To remove a host (node) from the cluster, either select the target host in the diagram (Diagram view) or click the Host tab and select that host in the table (Table view), and click the Remove Host link on the right of the Summary line. A dialog box appears to verify the action; click the OK button.
   1. The Prism web console displays a warning message that you need to reclaim the license after you have removed the node.
   2. Removing a host takes some time because data on that host must be migrated to other hosts before it can be removed from the cluster. You can monitor progress through the dashboard messages. Removing a host automatically removes all the disks in that host. Only one host can be removed at a time. If you want to remove multiple hosts, you must wait until the first host is removed completely before attempting to remove the next host.
   3. (Hyper-V only) Initiating a removal of a node running Hyper-V fails if the node is running as a part of a Hyper-V failover cluster and the following message is displayed.
      1. Node node id is a part of a Hyper-V failover cluster. Please drain all the roles, remove the node from the failover cluster and then mark the node for removal.
      2. If this message is displayed in either nCLI or in Web interface, cluster administrators must use the management tools provided by Microsoft such as Failover Cluster Manager to drain all the highly-available roles off the host and then remove the host from the failover cluster followed by removing the host from the AOS cluster.
   4. (ESXi only) VMware/cluster administrators must use the management tools provided by VMware such as vCenter to migrate all the VMs off the host and then remove the host from the vCenter cluster followed by removing the host from the AOS cluster.
      • Caution: Be sure to migrate VMs in an ESXi cluster prior to removing a host. Verify you have enough available compute capacity in the cluster (ESXi or Hyper-V) before actually migrating the VMs.
      • If you click Remove Host without first migrating the VMs, the VM’s data may be migrated without any notice and those VMs will lose service.
2. After a node is removed, it goes into an unconfigured state. You can add such a node back into the cluster through the expand cluster workflow.
3. To add a host into the metadata store, either select the target host in the diagram (Diagram view) or click the Host tab and select that host in the table (Table view), and click the Enable Metadata Store link on the right of the Summary line.
4. Each node includes a disk used for metadata storage, and AOS maintains a metadata store across these disks to ensure uninterrupted resiliency should a metadata disk fail. After such a failure, that node is taken out of the metadata store group and the cluster continues to operate seamlessly without it. Normally, the node is brought back into the metadata store automatically after the failed metadata disk is replaced. However, under certain (rare) circumstances this might not happen. If the node is ready but was not added back automatically, the alert message Node ready to be added to metadata store is displayed and the Enable Metadata Store link appears indicating you can add the node back into the metadata store manually.