First 3 things that come to mind are 1. Space management for the audit trail. If for every record DB2 needs to expand the space allocated for the audit trail table or for the tablespace then your bottleneck is there. The solution is to allocate more space for the tablespace and optimize extent size and/or buffer pools. I assume you use different tablespace for the audited table and the audit trail. 2. Check what options have been selected for the trigger. If you have email notification enabled and attempt to insert 750,000 records you will force the system to send 750,000 emails, which will surely kill the insert. 3. Transaction and rollback management. This could cause problems if the temporary space is not sufficient to handle large volumes of data (1,500,000 records – data + audit trail). : Thank you that is great! : Right now though we have noticed a performance issue need to understand : first. We are on the iSeries and have a pretty fast box but when we try to : mass update about 750,000 records in a file we have the update trigger on : it takes several times as long to run. This can not be simply due to the : trigger so we need to be sure this is correct. Any insight on wha tto look : for on the iSeries (AS/400) that may help? : Tom
|