Shutting down cluster nodes

Prerequisites

• If a node hosts NFS, block storage, or file storage services, unmount any iSCSI LUNs, NFS exports, or SMB exports on the client side before stopping the services. Failure to do so may result in data loss.

To gracefully shut down a cluster node

1. Put the node into maintenance mode and evacuate its workloads to other cluster nodes, as described in Performing node maintenance.

2. Release the node from the relevant services, as applicable:

   • To release a node from the compute cluster, refer to Releasing nodes from the compute cluster.
   • To release a node from the backup storage cluster, refer to Releasing nodes from backup storage.
   • To release a node from the S3 cluster, refer to Releasing nodes from object storage.
   • To release a node from the file storage cluster, refer to Releasing nodes from file storage.
   • To release a node from the block storage cluster, delete the iSCSI target on this node, as described in Managing targets.
   • To release a node from the NFS cluster, refer to Releasing nodes from the NFS cluster.
3. Verify the cluster health and ensure that no services are in a transitional state.
4. Release the disks from the node and wait for data migration to complete, as described in Releasing node disks.
5. Release the node from the storage cluster, as described in Releasing nodes from the storage cluster.
6. Remove the node from the infrastructure, as described in Removing unassigned nodes.

7. Power off the node once it has been fully released:

   • If you can access the node remotely, execute:

     # shutdown -h now
     

   • If you can access the node physically, briefly press the power button once.