The information in this article discusses possible reasons why
a job might experience poor performance.
| Insufficient memory |
|---|
| |
How to diagnose: |
View the properties of the job to determine which memory pool the job
is running in. Then view the properties of the memory pool in iSeries™ Navigator,
see Check memory pool use. A high
rate of faulting in a pool indicates that there is not enough memory in the
pool, or that too many jobs are in the pool competing for the memory. |
| Recovery: |
- Turn on the system tuner if you are not already using it. The system value
QPFRADJ automatically
adjusts memory pools and activity levels.
- If possible, manually tune the pool you are working with by increasing
the amount of memory in the pool or reducing the activity level for the memory
pool. You might also want to check the machine pool to verify that the amount
of memory being used is not affecting all jobs on the system.
|
| Activity level too low |
|---|
| |
How to diagnose: |
View the properties of the job to determine its status and which memory
pool the job is running in. If the job shows a status of Waiting
for activity level, then view the properties of the memory pool
in iSeries Navigator,
see Check memory pool use. A high
rate of transitions to the ineligible state in a pool indicates that too many
jobs in the pool are competing for the memory. |
| Recovery: |
- Turn on the system tuner if you are not already using it. The system value
QPFRADJ automatically
adjusts memory pools and activity levels.
- Manually tune the pool by increasing the activity level for the memory
pool.
|
| Insufficient CPU resource |
|---|
| |
How to diagnose: |
View the CPU % column for the job and other jobs in the Active Jobs
list of iSeries Navigator.
If the system is very busy, your job might not be getting enough CPU resource
to complete its work. |
| Recovery: |
- If possible, end or hold unnecessary work on the system.
- If a few jobs are CPU intensive, change the run priority of these jobs
(a higher run priority value equals a lower run priority for the job).
|
| Memory pool paging option |
|---|
| |
How to diagnose: |
If an application is disk intensive, if the CPU is under utilized and
if there is sufficient memory, the use of expert cache might be beneficial. |
| Recovery: |
The expert cache can be turned on in iSeries Navigator by changing the Paging
option for a shared memory pool to Calculated. The Paging option is located
on the Configuration tab of the memory pool's Properties page
and is only available on shared pools(not private pools). |
| Low job run priority |
|---|
| |
How to diagnose: |
View the job's properties to
determine the run priority of a job relative to other jobs on the system. |
| Recovery: |
If the job has a low run priority (higher number) relative to other
jobs and is not using much CPU because the higher priority (lower number)
jobs are using most of the CPU resource, you might need to increase the job's
run priority, see Job properties.
Also, on a system with high CPU utilization and a job with a low run priority,
setting the Dynamically adjust job priorities within priority bands (QDYNPTYSCD) and the
Dynamically adjust job priorities of interactive jobs (QDYNPTYADJ) system values might
be useful. |
For more information about performance, see Performance. If you want more information
about how to tune performance on your system, see Tune
performance.