Switchable devices

In high availability environments, IBM® i supports other switchable devices other than independent disk pools.

Prior to 6.1, IBM i only supported switching independent disk pool devices. When a switchover or failover occurs, the device cluster resource group (CRG) switches the independent disk pools devices from the primary node to the backup node. Starting in 6.1, other hardware devices might also switch with the independent disk pool devices. On that backup node, clustering ensures the independent disk pools report in with the same resource names, but the other non-independent disk pool devices may report in with different resource names.

Clustering will ensure that the resource names and underlying physical devices for non-independent disk pool devices controlled by a device CRG are the same on all nodes in the device domain. Information regarding the physical device, such as resource name and type, is saved from the node which owns the hardware and restored to the other nodes in the recovery domain. This is done when the configuration object for the device is included in a device CRG or when a node is added to the recovery domain. When you are adding a device entry or a node to the recovery domain, the resource name for the physical device must match on every node in the recovery domain or these operations will fail. You can either do this manually or automatically by using cluster administrative domain to keep resource names of these devices consistent across nodes in the recovery domain.