I have some schedules report in Cognos 8.4.1 running on AIX ( with an Oracle content store and Progress OpenEdge reporting datbase) which fail intermittently.
When they fail I get the error RQP-DEF-0112 Query execution time exceeds the 1200 second limit in the cogserver.log.
I know I can increase this limit in model but my confusion is to why the reports run 98% of the time and run with in 180 seconds ( well with in the limit) but then on some occasions fail having taken nearly 10 times longer to execute. There is nothing out of the ordinary happening on the reporting database or on the Cognos server ( the report is scheduled for the same time each week). I can then manually run the report immediately after the schedule has failed and it succeeds.
Has anyone else experienced this kind of behavior ?
Thanks
Mike
Increase the timeout. See here:
http://pic.dhe.ibm.com/infocenter/cbi/v10r1m1/index.jsp?topic=%2Fcom.ibm.swg.ba.cognos.c8pp_inst.10.1.1.doc%2Ft_soap_fault_trbls.html
Is there criteria being entered in to the report at run time that is causing the query to take extra long?
For example, a report that summarizes sales transactions may have a Date range entered on the prompts page. Most users run the report for a week or two in the date range. One specific user decides they want the report for the entire prior fiscal year, so they use a very large date range. The query takes a huge amount of time to execute because of the large amounts of data involved.
You can check the report history for completed reports to see what criteria was entered in on the prompts page.