HP StorageWorks XP Remote Web Console Software - Manual (Page 107)

TrueCopy for z/OS user guide 107
without any increase in service or disconnect time. The asynchronous R-VOL updates are performed
completely independent of all host I/O processes at the M-VOLs and there are no restrictions on
subsequent read/write operations to TC390A M-VOLs. The only performance concerns for TC390A are
ensuring that adequate cache resources are available for sidefiles that are used to store the recordsets
at both the MCUs and RCUs, and ensuring that sufficient ESCON paths are defined for copy
operations.
In addition to the TC390 update copy mode, several other factors can also affect
XP128/XP1024/XP10000/XP12000 performance. You can optimize both the TC390 operations and the
I/O performance of the disk arrays by analyzing workloads and addressing system-level conditions (for
example, number of ESCON paths) that can affect disk array performance. You can also control the
impact of TC390 operations on disk array performance by selecting the appropriate RCU options for each
MCU (see ” Registering an RCU (Add RCU) ” on page 81) and the appropriate update copy mode
(synchronous or asynchronous) and pair options (see ” Creating TrueCopy Volume Pairs (Add Pair) ” on
page 115) for each TC390 pair. In addition, you can upgrade the TC390 hardware components and/or
adjust the configuration of the components to improve disk array performance under a wide range of
operating conditions. The following table lists some of the conditions that affect disk array performance
and provides recommendations for addressing these conditions.
Table 21 Optimizing TrueCopy operations and XP128/XP1024/XP10000/XP12000 performance
Condition Description Recommendation(s)
Write-intensive
workloads
Write-intensive workloads, such as SPOOL
volumes and database logging volumes,
can have a significant impact on disk
array I/O response times.
Spread write-intensive data across several
volumes to minimize queuing. Also
consider increasing the duplex write line
(DWL) of the disk array using Performance
Manager.
Large block size Workloads with large write block sizes,
such as DB2® deferred writes, can impact
performance.
Spread workloads with large write block
sizes across several volumes.
High host channel
demand
The demand on the MCU’s host channels
can affect performance.
Spread the workload across several disk
arrays to use additional channels.
Sequential write
operations
TC390 operations can have a negative
impact on workloads with a high
percentage of sequential write operations,
such as batch processing operations (for
example, dump/restore or sort
operations).
Avoid performing restore operations to
volumes that belong to TC390 pairs.
Instead, restore data to a scratch volume
and then establish the TC390 volume pair.
Cache size Large cache size improves read hit
performance, which allows more disk
array resources to be devoted to write
operations. The resulting performance
improvement can offset some or all of the
performance loss due to the TC390 remote
copy operations.
TC390A and HXRC require additional
cache for sidefile data. Insufficient cache
resources can result in command retries,
SCP notifications, and puncture conditions.
Consider increasing the cache size of the
TC390 disk arrays to handle TC390A and
HXRC sidefile operations and to improve
overall disk array performance. For best
results, the cache and NVS capacity of the
main and remote disk arrays should be the
same (for TC390A the RCU sidefile
requirements are 2x that of the MCU) to
enable the remote site to function
adequately during disaster recovery.