Rebooting an IBM SP2 cluster

EG, the ESMF.

PMR #68524-227

drain loadleveler

All on the CWS unless otherwise noted:

Equiesce -p all if doing all the nodes, but not for rebooting multiple nodes

Efence <hostname> or <node num>

On each node individually:
shutdown -F or shutdown -Fr

Or alternatively, from the CWS:
cshutdown <node num> sep by commas or 3-7 for nodes 3 thru 7.
Also a -r flag.

Estart after all nodes are back up

If a boot problem, go to NIM maintenance shell.

splstdata -b gives node numbers

If the SP2 switch doesn't come back up right (you cannot ping one or
more of the "s" addresses), then please see:

https://stromberg.dnsalias.org/~strombrg/SPswitch.html