Re: Oracle performance and swapping on Tru 64

Thanks for the swift reply. The statspack reports all indicate that
'free buffer waits' are a major problem, which is what the 'internal
write cache' tuning recommendation is supposed to address. I think that this is caused by the commit on every change. Another issue - on the ever growing list - is that there's only 3 disks allocated for use by Oracle on the server and the way the filesystems have been setup means that there's contention between online redo log and archive redo log generation. Unfortunately, the server is so old that getting new disks for it is a problem, apparently there was a disk crash a couple of years ago and only 2 could be sourced, so spreading the load over more disks doesn't look like an option either. I've just had a call from one of the guys in the UNIX performance team to say that some of the values being reported by the disks for response time etc. are
'completely unrealistic and spurious', leading him to believe that we
may have disk problems. I'll post the iostat output from the same period as the vmstat output posted earlier, just in case there's anything you can spot.
Thanks again,
  Rob
21:30

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 01167 36 2 0 1163 38 0 0 574 62 2 32 6 60    0 0 879 26 26 3 2056 52 0 0 594 67 2 35 13 50    0 02790 59 0 0 734 22 0 0 664 68 2 28 10 59    0 0 463 17 6 1 461 14 0 0 618 66 2 12 6 81    0 0 18 2 14 2 30 3 0 0 614 66 1 0 8 92 21:35

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 01213 34 4 1 1289 38 0 0 743 75 1 22 5 72    0 02854 56 0 0 814 26 0 0 745 74 2 13 6 78    0 01182 34 26 3 1217 38 0 0 767 75 1 21 9 69    0 0 406 12 0 0 454 17 0 0 713 72 1 7 2 90    0 02841 66 0 0 794 27 0 0 702 72 3 13 6 78 21:40

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 0 16 2 24 3 71 7 0 0 674 71 1 0 6 92    0 0 10 1 0 0 62 6 0 0 698 73 0 0 1 99    0 0 262 11 14 0 344 16 0 0 680 71 2 6 2 90    0 01830 43 14 1 1018 33 0 0 734 75 1 18 10 71    0 02239 48 0 0 1054 33 0 0 766 79 1 17 4 78 21:45

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 0 17 2 26 3 78 7 0 0 690 72 1 0 7 92    0 0 10 1 0 0 42 5 0 0 712 74 1 0 2 97    0 02206 41 0 0 878 32 0 0 634 71 4 2 5 89    0 0 971 28 16 2 3257 89 0 0 789 72 6 17 11 66    0 01185 34 0 0 1224 36 0 0 854 77 6 21 6 68 21:50

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 0 17 2 27 3 71 7 0 0 671 71 4 1 7 88    0 0 11 1 0 0 48 5 0 0 582 64 0 0 1 99    0 0 12 2 0 0 61 6 0 0 666 69 0 0 1 99    0 0 14 2 9 1 70 7 0 0 670 70 1 1 6 92    0 0 10 1 0 0 66 7 0 0 685 71 0 0 1 99 21:55

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 0 17 2 25 3 68 7 0 0 709 74 1 1 7 91    0 0 10 1 0 0 33 3 0 0 610 70 0 0 1 99    0 0 10 1 0 0 54 6 0 0 668 70 0 0 1 99    0 0 14 2 10 1 75 7 0 0 693 73 1 1 7 92    0 0 244 9 0 0 319 14 0 0 660 70 1 5 1 93 22:00

      tty re1 re8 re9 re16 re17 cpu  tin tout bps tps bps tps bps tps bps tps bps tps us ni sy id    0 4 405 12 95 3 520 18 151 2 467 22 10 7 4 79    0 02322 48 38 3 1583 42 0 0 814 87 2 6 11 81    0 0 10 1 0 0 137 15 0 0 857 89 0 0 2 97    0 0 14 2 0 0 150 16 0 0 872 92 0 0 2 97    0 0 18 2 14 2 130 16 0 0 854 91 1 1 8 90    0 0 10 1 0 0 106 13 0 0 862 92 1 0 1 98 Received on Mon Mar 20 2006 - 09:08:07 CST