Re: Perf. problems with multiple users

>> Most sessions are only active for a short period of time,
it's only my session that can be active for up to 5 minutes. During these five minutes the system is unworkable for all other users. <<

If the system works fine before your query runs and while your query runs other sessions basically freeze up you must be consuming all available resources or a latch somewhere that is halting work. Try running the bstat/estat jobs around your job so that you can find this resource.

You may just be exhausting the dictionary cache or library cache and need a bigger shared pool. Perhaps you need to increase the size of the buffer pool or sub-divide it into a keep and regular pool. The bstat/estat statistics should point us at the right area.

And even though you do not need results in less than 5 minutes the rdbms may need you to tune the query in order to be able to provide resouces to other jobs. The only time I have seen the rdbms freeze up like you are describing except when error messages appear in the alert.log is when an update was being done which caused all redo logs to fill and the archive was not complete on the first log before this happened.

I think you are going to need to post the bstat/estat statistics or at least selected statistis for the offending session before anyone can identify the problem.

Got questions? Get answers over the phone at Keen.com. Up to 100 minutes free!
http://www.keen.com Received on Fri Jul 14 2000 - 00:00:00 CDT