HP XP Racks - Manual (Page 57)

Recovering from secondary volume failure on the RCU
When this failure occurs, automatic system failover switches host I/O to the primary volume on
the RCU. Recovering the volume from this failure involves releasing the pair, clearing the failure,
recreating the pair, and restarting host I/O. Unlike the other types of pair volume failures, host
I/O is not interrupted and failover does not occur.
NOTE: If the failed volume is an external volume, contact the HP Support Center for assistance
clearing the failure.
1. Using multipath software, make sure that the primary storage system is receiving host I/O.
2. Using multipath software, vary the non-owner path offline.
3. On the primary storage system, run the Pairsplit-S RAID Manager command to release
the pair.
4. Clear the failure in the secondary storage system S-VOL.
5. If using RAID Manager, on the primary storage system, run the paircreate RAID Manager
command to create the ESAM pair.
6. If using RWC, on the primary storage system create the ESAM pair using the Paircreate(ESAM)
dialog box.
7. On the primary storage system, create the pair (Paircreate(ESAM)).
8. On both systems, make sure that Type shows ESAM.
9. Using multipath software, vary the non-owner path online.
10. Restart host I/O.
Recovery from quorum disk failure
There are two basic methods for recovering from quorum disk failure. In some cases, the disk must
be replaced. In other cases, you can recover from the failure by resynchronizing the pair volume
that is connected to the disk.
•
“Replacement of quorum disks” (page 57)
•
“Recovery from failures using resynchronization” (page 62)
Replacement of quorum disks
Replacement of a quorum disk is done when the disk fails and the external storage system that
contains the disk cannot be recovered by resynchronizing the volume pair connected to the disk.
The procedure you use varies depending which system (primary or secondary) is receiving host
I/O.
The procedures are:
•
“Replacing a quorum disk when the MCU is receiving host I/O” (page 57)
•
“Replacing a quorum disk when the RCU is receiving host I/O” (page 58)
NOTE: You can use the replacement procedures to replace a disk that is connected to one or
more volume pairs.
Replacing a quorum disk when the MCU is receiving host I/O
Replacing the disk involves preparing the systems and the volume pair (or pairs) connected to the
disk so that you can safely remove the failed disk and replace it. After it is replaced, you reconnect
the systems to the disk, add the disk ID, and re-create the volume pair (or pairs).
When you replace a disk, data on the disk is erased and you cannot continue using ESAM.
CAUTION: To prevent host I/O from stopping, make sure you complete the steps exactly as they
are listed in the procedure.
Recovery from quorum disk failure 57