One node of a 2-node cluster is shut down to replace a NIC. All RDB rings fall out of quorum. To diagnose the problem, what should the administrator do first?

Disconnecting and reconnecting an InfiniBand cable is a physical action that should only follow logical diagnostics, not the first troubleshooting step.

'storage failover show' reveals failover state but does not directly address RDB quorum loss, which is a cluster-level issue best diagnosed at the HA layer first.

'cluster ha show' reveals whether cluster HA is enabled on the surviving node, which directly affects quorum behavior in a 2-node cluster. In a 2-node cluster, if HA is not enabled or is misconfigured, the remaining node may not be able to maintain quorum on its own, and this command is the fastest first diagnostic step to confirm the HA state before deeper investigation.

Verifying the cluster switch is a valid step but comes after confirming the logical cluster HA configuration, which is quicker to check and more directly tied to quorum.