HP 3PAR T-Class Installation and Deinstallation Guide

D Performing Node Rescue
This appendix describes the following methods for recovering a node that has been shut down or
recently added to the system. It is possible to “rescue” a node with the following methods:
• Nod to node rescue: this enables rescue of a node by manually connecting it to a working
node.
• CLI node rescue: this launches the startnoderescue command from a command line interface
(CLI) to trigger detection and rescue of the blank or shut-down node.
Node-to-Node CLI Rescue
The method enables you to rebuild a node disk (performing node-rescue) using a working node
and the CLI command startnoderescue.
Requirements include:
• Two Ethernet connections to the same network.
• One node in the storage system (the rescuer) node must have a functioning network connection
and the system must be able to process HP 3PAR OS CLI commands.
• The node-to-be-rescued (the target) node must have an Ethernet cable connected to the
Administrative network port (E0) and be connected to the same network as the rescuer-node
in the storage system.
Figure 117 Node-to-Node Rescue Port Connections
• The node to be rescued must be at the BIOS prompt. If the node is running it must be halted
using shutdownnode halt <node_ID>.
To rescue a node using another functioning rescuer node, do the following:
1. Start an spmaint session.
2. From SPMAINT, enter the Interactive CLI shell for the storage system.
3. Enter the CLI command shownet to determine which node has the ACTIVE Ethernet connection.
This is the rescuer. For example, node-1 has the active network connection in the following
output:
cli% shownet
IP Address Netmask/PrefixLen Nodes Active
192.168.56.206 255.255.255.0 01 1
4. If necessary, remove the Ethernet cable from port E0 of another node in the system.
5. Connect an Ethernet cable (or the one temporarily removed in step-3) to port E0 of the
target-node.
6. Enter the CLI command startnoderescue -node target_node, where target_node
is the Node ID (slot number) of the node being rescued.
202 Performing Node Rescue