I think these set of question will allow you to find the issue when gathering information from end users about a report in the Cognos environment that is not performing as expected.
Hope this helps many. Please enjoy:
- Has the report always been slow or did it recently become slow?
If it recently became slow, can you identify an event that occurred just before the report began to run slowly? Events could include changes to configuration settings, changes to tuning settings, a recent upgrade where your previous settings have not been applied, an introduction of firewalls or proxies, changes to existing firewalls or proxies, changes to virus scans on temp directories, or temporary table space restrictions on the database. This event could have caused the change in report performance.
- Is the performance slow for all reports or just one report?
If all reports are slow, the issue may be due to your environment or database. If all reports from a specific package are slow, the issue may due to the model design. If just one report is slow, the issue may be due to a specific report element.
- How many queries does your report contain?
The number of queries on the report will proportionally affect the report execution time.
Does the report run slowly for everyone, or just for one user?
If the report runs slowly for just one user, the issue may be due to something in that user's environment, such as virus scanning, page file size or location settings, or their location on the network.
- Is the report burst or run often by many people?
If many people are running the same report at the same time, you may need to scale your environment or consider using dispatcher routing rules to direct all requests for a specific package or group of users to a specific server or server group.
- Do your queries require local processing?
The following report elements require local processing: crosstabs and charts, master relationships, unions or joins, multiple fact queries, bursting, and non-vendor specific functions. Local processing requires the IBM Cognos server to compute operations on the result set returned by the database, which can impact the SQL execution time.
- Have you reviewed the logs in the c10_location/logs directory and the audit logs?
They may help you identify the source of the problem. Monitoring your processes, such as the Java and Business Intelligence bus processes could also identify excessive memory use.
- Have you recently upgraded?
Ensure that any tuning settings that were applied to your previous installation are applied to the new environment. Ensure that your models have been verified, upgraded, and republished. Verify that the IBM Cognos Framework Manager governor that allows enhanced model portability at runtime is not enabled. Depending on your upgrade method, you may also need to open and save the reports again after upgrading.
