HP Storageworks HSG80 Manuale Utente

Platform-Specific Issues

12

If you are unable to mount a previously configured file system after a failover or failback, run 
File System Check on the logical volume and retry the mount.

Example: FSCK -O full /dev/vg0#/lvol#

Remote copy sets (RCSs) at the target site will acquire the world wide name of their RCSs at 
the initiator site when a failover occurs. HP recommends that you not configure the RCS on 
the HP-UX host servers at the target site until after a failover.

Remedial kits that you may require are available at:

If during the failover procedure you are unable to stop all applications and dismount all units 
that are part of a remote copy set, you must reboot hosts prior to failback. Doing this prevents 
possible data corruption caused by writing stale data from host cache to the units after 
failback.

Extreme care must be used when creating and managing remote copy sets on an ATM link. If 

these links are overstressed (for example, an excessive number of remote copy sets created for a 

given environment), unexpected behavior may result, such as loss of remote copy sets and link 

failures.

HP recommends that no attempt be made to access the special device files during a fabric 
failover. Running any commands against the block or raw devices in /dev/disk or /dev/rdisk is 
highly discouraged. In rare instances, running the 

FILE

 command against raw or block 

devices during a fabric failover has prevented the surviving path from coming online following 
failure of the original path.