beautypg.com

Considerations for virtual disk processes, Pool catalog placement, Pool cpu assignment – HP Integrity NonStop J-Series User Manual

Page 54: Gathering pool statistics, Reporting disk full ems events

background image

Configuring and Managing SMF Processes

HP NonStop Storage Management Foundation User's Guide523562-007

3-4

Considerations for Virtual Disk Processes

Pool Catalog Placement

A pool catalog must be on a mirrored, audited volume. The default location is
$SYSTEM. Because a pool catalog is not large or heavily used, all pool catalogs can
share a disk; however, because of naming restrictions, a maximum of 100 pool
catalogs can be placed on the same volume. You can place pool catalogs on multiple
volumes by specifying the catalog location when the pool is configured.

Pool CPU Assignment

You assign the pool process a primary and backup CPU pair when you configure it. If
you are pooling by CPU, you should assign the pool processes to that CPU.
Otherwise, you can spread the pool processes across all CPUs to balance the CPU
and memory load. Choose the backup CPU locations to spread the SMF processes
across multiple CPUs in the event that the primary CPU fails. Otherwise, excessive
memory pressure and CPU load could occur.

If you are pooling by CPU, place the pool process’ catalogs on the same CPU as the
disk process; otherwise, distribute the pool processes catalog to balance CPU memory
load.

Gathering Pool Statistics

SMF collects statistics such as the amount of disk space used and the number of files
allocated and uses them to select volumes when creating new files. You control the
interval at which these statistics are gathered with the UPDATESTATS storage pool
attribute. The default value is one minute. If you have a pool whose data is very static
and whose disks are not full, you should use a larger value, which will improve
performance. Statistics overhead will be greatest for pools with many disks and virtual
disk processes.

Reporting Disk Full EMS Events

Each pool can be configured with a disk full percentage threshold, which causes an
EMS event to be generated when any disk in the pool grows beyond the threshold.
Additional EMS events will be generated at periodic intervals as the disk continues to
grow. This information is used by the Automated Storage Manager (ASM) and can also
be used by your own EMS event filters.

By default, this feature is disabled. Consider enabling it for pools with nearly full disks
or with rapidly growing files. Refer to

Appendix B, COUP and PUP Interfaces

for

information on how to configure these thresholds.

Considerations for Virtual Disk Processes

Configuring the virtual disks is the most critical aspect of SMF configuration planning,
as the virtual disk process is performance critical, and there are usually many more
virtual disk processes than pool processes. The major parameters to be selected

This manual is related to the following products: