Warning

Before adding previously used drives from non-SWARM systems for reuse in a Swarm cluster, you MUST perform a ‘disk wipe’ of those drives to remove any leftover partitioning or other artifacts. See here for information about preparing a repurposed drive.

Physical volumes can be moved between nodes as necessary to address hardware failures or other constraints as determined by an administrator.

The remaining cluster nodes immediately guarantee the correct number of replicas exist for all objects in the cluster when a volume goes offline because of a volume failure, node failure, or node shutdown. The remaining nodes recognize the volume has returned to service and cease efforts to replicate the volume's data if a volume or node returns to the cluster during this procedure and before the 14-day time limit.

Important

Verify the node has enough RAM to handle volumes when adding or relocating volumes to a node. A node may be unable to mount some of the volumes if not.

Warning

If the Swarm drives were formatted for use with a RAID controller, do not move Swarm disks between disk array controller types after they are formatted by Swarm. Each RAID controller reports available disk space to Swarm that matches the controller. Many controllers claim the last section of the disk, reducing the total available space. A new controller may claim additional disk space not reported to Swarm if you switch your disks with another controller, so Swarm may attempt to write data to non-existing space, generating I/O errors.

Moving Clusters

Volumes can also be moved to nodes in a different cluster. The objects on the volume become part of the new cluster and are checked for the correct constraints within the context of the new cluster when this occurs.

Swarm Documentation

Moving Volumes between Nodes

Warning

Important

Warning

Moving Clusters