By default, CVM enables a node to join the cluster as long as the node has access to the disks through at least one node in the cluster. Typically, the node accesses the disks through the master node. A shared disk group can also be imported if some nodes cannot access all of the disks. During a node join operation or a disk group import operation, if CVM detects that the node does not have access to the underlying storage for the volume, the volume is placed in the LDISABLED state. The storage connectivity in such a situation is asymmetric, meaning that not all nodes have the same access to all the disks in the shared disk group.

This behavior is enabled by default. You can set the storage connectivity policy to resilient to disable this functionality. The configuration of this feature is independent from the detach policy and the I/O shipping policy. However, the disk group version and the cluster protocol version must be a level to support the behavior.

When the storage connectivity is set to resilient, Cluster Volume Manager (CVM) requires that a node joining a cluster must have access to all the disks in all shared disk groups. This requirement applies also to a node that was removed from the cluster because of an I/O failure. The node cannot rejoin the cluster until the disk access is restored. Similarly, a shared disk group cannot be imported unless all of the nodes in the cluster can access all of the disks in the disk group.

See Controlling the CVM tolerance to storage disconnectivity .