Re: Peformance degradation

Hi Craig, Sounds very interesting, your problem. Is it possibe that you have defined a clustered index on a very dynamic table? The effect being that many deletions cause holes to appear, these holes could be very costly. More likely is a need to do regular update statistics. I would be interested in seeing some of your tbconfig, in particular I'd like to know how long your checkpoints are taking. You can guesswork this from the tbstat -m output (assume that the checkpoints start at n minute intervals on the stroke). When you say reorganization what do mean?, have you unloaded dropped and recreated the tables involved ? Have you Exported/Imported the database?. Use the series of tbchecks to examine the database... tbcheck -cc, tbcheck -cr, tbcheck -ce, tbcheck -cI <dbname> and tbcheck -cD. These should take you some time. When you've finished that, check that the contents of your systables is intact, use select tabname, npused from systables where npused < 0. It has been known to throw the optimizer. Look at the output from the SET EXPLAIN ON statement on offending/inefficient sql statements. Maybe you need to seperate your logical logs from the rootdbs. I don't believe that the problem is with striping. When you say that you're sharing the chunk with other non-Informix data do you mean this beast is on cooked files? There are so many possibilities here. For a lot of folks in C.D.I. your performance problems are like a starting gun, its like...time to go to work. Cancel your weekend, grab the fine manuals, the Suto, the Lumbley and the FAQ, if its possible, get the system to yourself and go for it. Hope This helps, Tom Hogarty <STANDARD and EXCEPTIONAL DISCLAIMERS APPLY> Pyramid Technology Corp -- Tom Hogarty