Posted by tecwrg on 2/16/2011 11:31:00 AM (view original):
We're seeing an incrreased amount of lock escalation on a particular table during the production batch cycle for one of our applications. This in turn is leading to an increased number of -911 lock timesouts and/or deadlocks. I've tried dumping the dynamic SQL cache and looking at the statements against this particular table, and I've also run the db2advis command against the cached workload looking for indexing opportunities, but nothing is jumping out at me as the cause of the problem.
Any suggestions? Should we just keep increasing the size of the locklist and hope for the best?
On a scale of 1 to 10, with 1 being "who gives a sh*t" and 10 being "that's awesome tec, way to go!", how pleased are you to know that after a week of mind-numbing analysis, I've finally identified and resolved the problem?