HP 3PAR AIX and IBM Virtual I/O Server Implementation Guide

11 AIX Client Path Failure Detection and Recovery
This chapter explains how to set up the AIX client LPARs to enable client automatic path failure
detection and recovery in the event an IBM Virtual I/O Server goes down for any reason.
AIX Client Automatic Path Failure Detection and Recovery
When one of the VIO servers goes down for any reason, the vscsi path coming from that server
shows as failed with the lspath command when executed from the virtualized AIX client.
# lspath
Failed hdisk0 vscsi0
Enabled hdisk0 vscsi1
Setting Auto Path Failure Detection and Recovery
Even though the VIO server comes back as available, the lspath command will still show one of
the paths as missing. For automatic path failure detection and recovery on the AIX client, set the
attributes hcheck_interval and hcheck_mode to 60 and nonactive, respectively. This will
cause a path failure to be detected automatically and recovered automatically once the path has
returned.
chdev -l hdiskN -a hcheck_interval=60 -P
chdev -l hdiskN -a hcheck_mode=nonactive -P
This task must be performed for each hdisk on the AIX client.
The AIX client will need to be rebooted for the 'hcheck_interval' attribute changes to take
effect.
As new disks are added to the AIX client, the command to set the 'hckeck_interval' must be
performed for each new hdisk added. The same is true if an hdisk is removed (using the rmdev
-l hdiskN -d command) from the AIX client and added again later.
For further information regarding path failure detection and recovery, see the IBM manual on the
website:
www.ibm.com
AIX Client Automatic Path Failure Detection and Recovery 49