Re: 7.22.UC1X2 Performance tuning questions....

This system is huge. I don't know if you really need 300000 buffers or not, but let's assume that you do. The first thing that I would do is to set KAIO. With 10.10+ Informix is doing "tag" io for KAIO. This means that we are no longer doing select() to see if the IO is complete. This has improved KAIO performance over HPUX 9.x quite a bit. It will require that residency to be turned on, but if you have enough memory to handle this large of a system, this should not be a factor. Next - I'd consider lowering the LRU_MAX to 2. I realize that a LRU_MAX of 5 might seem quite low, but with 5% of 300000 is still 15000 pages and that's a lot of work to actually be doing at a checkpoint. Realize that when a page is written to disk because the LRU_MAX/LRU_MIN has been reached, it is not removed from the buffer, it's simply copied to disk. This means that by having some of the writes done between checkpoints will reduce the length of time that a checkpoint will take. The "performance advantage" of not writting the page to disk immediatly after an update of that page is only realized if there is another update to the same physical page before the checkpoint occurs. Next -- make sure that your checkpoints are really only occuring every 10 minutes. If they are occuring any earlier, then your physical log file is too small. You need to increase it. I'm guessing that you are not running client programs on the same box as the database because you have 8 processors with 7 cpuvps. Make sure that your net poll thread is running on the CPUvp and not on a NET vp. Subject: 7.22.UC1X2 Performance tuning questions.... From: tony.edwards@bellsouth.co.nz (tony edwards) Date: 2 Apr 1997 10:25:30 -0500 Message-ID: <5httpa$nsl@cssun.mathcs.emory.edu> Team, We have recently upgraded our HP-UX 9.04 / 5.07 to HP-UX 10.20 / 7.22.UC1X2. Here is some information about the system configuration...... Database size 50 GB, chunks 43, memory 3 GB, disk nike arrays 120GB, cpus 8 NUMCPUS 7 LOCKS 600000 BUFFERS 300000 CKPTINTVL 600 NUMAIOVPS 20 PHYSBUFF 400 LOGBUFF 400 LOGSMAX 165 CLEANERS 45 SHMVIRTSIZE 1000000 SHMADD 200000 LRU_MAX_DIRTY 5 LRU_MIN_DIRTY 1 LRUS 45 STACKSIZE 32 OPTCOMPIND 2 NOAGE 0 We don't have KAIO set because processor affinity is not supported on HP. Problems we are having.... 1) A checkpoint is taking 16-20 seconds (Stats based on interval of 300). 2) Ready to run can be over 60 processes. 3) The oninit processes never go below 156 for a priority and can be as high as 268. 4) System not using indexes correctly - Have run update statistics medium for all tables, update statistics high on first column of each index and run update statistics low on all other index fields. Can any one advise me on their update statistics strategie ? (This system is 24x7) 5) Our LRU writes are 60% and the chunk writes are 40 % of the onstat -F command. 6) If we changed our optcompind to zero would we need to run so many update stats? I look forward to any suggestions you can provide... Cheers, Tony. Madison Pruet