HP XP Continuous Access Software - Manual (Page 266)

266 HP StorageWorks Disk Array XP RAID Manager: User’s Guide
horctakeover
(swap-takeover)
When executing horctakeover on a standby server manually, I/O
activity on the servers (for the pertinent CA volumes) must be stopped.
Host machines that can own
opposite sides of a CA pair Host machines must be running the same operating system and the
same architecture.
New RM installations After a new host system has been constructed, a RM failure to start can
occur due to an improper environmental setting or an inaccurate
configuration definition file. Use the RM activation log file for RM
error definitions.
Host failure If a failure occurs on host A (with application failover software
installed), host B detects the failure and issues a takeover command
making the secondary volumes usable. If the secondary volumes can
continue processing, host B takes over processing from host A.
While host B continues to process (swap-takeover) the command, the
volumes are swapped so that the secondary volumes become the
primary volumes and the primary volumes become the secondary
volumes. When host A has recovered from the failure, it can take back
ownership and control through another swap-takeover command.
Secondary volume failure If the primary volume detects a failure in the secondary volume, pair
writing is suspended. The primary volume changes the paired volume
status to PSUE. (The fence level determines whether host A continues
processing, that is, writing, or host B takes over from host A.) RM
detects the change in status and sends a message to the syslog.
If host A had initiated a monitoring command, a message appears on
host A. When the secondary volume recovers, host A updates the
S-VOL data by running the pairsplit –S, paircreate –vl, or pairresync
command.
Startup failure When the P-VOL server boots up, the secondary volume can be
updated. If the secondary volume is used by the LVM, the volume
group of the LVM must be deactivated. The secondary volume must
only be mounted to a host when the volume is in PSUS state or in
SMPL mode. The secondary volume must not be mounted
automatically in any host boot sequence.
(continued)
Error Solution