RE: Performance Issues

A couple of notes. I do not have any experience running informix on NT, just UNIX. Read up on all of these parameters before setting any of them. If you have any questions about them, post them to the newsgroup or email me. 2000 BUFFERS is way to low. Is the database the only thing running on this server? If so, I would guess you should run no less than 75000 BUFFERS, if not more. Increasing the BUFFERS to this large of a number will cause a bunch of other tuning parameters to be neccesary. You might start out at 50,000 with plans to keep increasing as long as you see a benefit. Maybe going no higher than 125,000... Other parameters which need to be set at the same time. LRUs I usually set at BUFFERS/1000. With no more that 128(the max) CLEANERS=LRUS is a good starting point for cleaners To control the length of checkpoints I would start with the following parameters for LRU_MAX_DIRTY and LRU_MIN_DIRTY. Watch your checkpoints, and if they are to long drop these numbers lower. LRU_MAX_DIRTY 5 LRU_MIN_DIRTY 3 Definition of parameters discussed. BUFFERS specifies the maximum number of shared-memory buffers that the database server user threads have available for disk I/O on behalf of client applications. LRUS Specifies the number of LRU queues for the shared-memory buffer pool CLEANERS Specifies the number of page-cleaner threads that the database server is to run LRU_MAX_DIRTY Specifies the percentage of modified pages in the LRU queues that flags page cleaning to start LRU_MIN_DIRTY Specifies the percentage of modified pages in the LRU queues that flags page cleaning to stop Hope this helps, Will >===== Original Message From "Kirk Davis" <kirk@kirkdavis.co.uk> ===== >As my earlier post, we have just upgraded our Informix Dynamic Server 7.3 >hardware to a 4 * 550mhz Xeon with 1024mb RAM.. The platform was previously >a single 200mhz >Pentium wth 256mb RAM. > >As suggested, I have set the following to optimise performance: > >NUMCPUVPS 3 >MULTIPROCESSOR 1 >SINGLE_CPU_VP 0> >As well as UPDATE STATISTICS, and increase the buffers to 2000, but the >system does not seem to be running much faster than before, and I have run >out of ideas. > >Can anybody help??? > >Thanks in advance, >Kirk ------------------------------------------------------------ This e-mail has been sent to you courtesy of OperaMail, as a free service from Opera Software, makers of the award-winning Web Browser, Opera. Visit us at http://www.opera.com/ or our portal at: http://www.myopera.com/ Your free e-mail account is waiting at: http://www.operamail.com/ ------------------------------------------------------------