Maximum Number Of Pairs For The Xp10000/Xp12000 - HP StorageWorks XP10000 User Manual

Table 3 BC operational requirements for XP12000 and XP10000 (continued)
Parameter
Dynamic sparing and
automatic correction copy
Physical device (PDEV)
maintenance
Logical device (LDEV)
maintenance
Cache maintenance
Point-in-time backup
Failures
Maintenance/upgrade
1. Maximum number of pairs per disk array depends on installed firmware version.

Maximum number of pairs for the XP10000/XP12000

When you create BC pairs, resources called differential tables are required. The number of differential
tables available in one disk array depends on the array's amount of shared memory for differential tables.
• You can use 13,652 differential tables if no additional shared memory for differential tables is
installed.
• You can use 30,718 differential tables if additional shared memory for differential tables is installed.
NOTE: Contact your HP account support representative for more information about adding additional
shared memory.
NOTE: The maximum number of pairs you can create is 8,192, no matter how much shared memory for
differential tables is available. If a command device is installed in order to use RAID Manager, the
maximum number of pairs is 8,191 (when configuring pairs of 1 P-VOL: 1 S-VOL ratio).
To calculate the maximum number of BC pairs, you must calculate how many differential tables are
required to create BC pairs, and compare the result with the number of differential tables in the disk array.
The following applications also use differential tables:
• ShadowImage for z/OS
• Hitachi FlashCopy Mirroring
Requirement
If a failure occurs requiring the disk array to utilize dynamic sparing or automatic
correction copy, the status of the paired volumes associated with the failed physical
device is not affected.
Maintenance of a PDEV that is not used by BC does not affect BC operations or the
status of BC pairs. If a PDEV contains LDEVs used by BC, you can maintain that PDEV
independently of the BC operations and pair status.
If the status of a BC pair is COPY(PD), PAIR, PSUS, PSUS(SP), COPY(SP), COPY(RS), or
COPY(RS-R), maintenance for the LDEVs used by that pair is restricted. Maintenance is
also restricted for LDEVs with the Reserve attribute.
To maintain these LDEVs, suspend or delete the BC pair, or reset (unreserve) the
Reserve attribute.
If disk array cache maintenance is performed during a period of high I/O usage, one
or more BC pairs may be suspended. Reduce the I/O load before cache
maintenance.
For pairs in PAIR status, host I/Os are copied to S-VOLs asynchronously. To
synchronize a pair and access the S-VOL as soon as possible, vary the P-VOL offline
and split the pair before using the S-VOL. If the P-VOL cannot be taken offline, a P.I.T.
backup still occurs if a pairsplit is issued. When all data on the P-VOL (up to the
pairsplit time) is copied to the S-VOL, the pair is split and the S-VOL becomes
available for use (R/W). The operation may take longer to complete if high write
activity from the host to the P-VOL is occurring. A P.I.T. backup can also be performed
by issuing a pairsplit to P-VOLs in the SMPL (simplex) state.
When a failure of any kind prevents a BC copy operation from completing, the disk
array suspends the pair. If an LDEV failure occurs, the disk array suspends the pair. If
a PDEV failure occurs, BC pair status is not affected because of the RAID architecture,
but copy pace may be degraded.
Under certain conditions, BC volume pairs may need to be re-established.
Business Copy XP user guide for the XP10000/XP12000 15