Hi Guys! In following this thread, I thought I'd check out customer experience with a stock TSM server they have.
q db f=d showed the DB hit rate at 85% q opt showed the bufpoolsize at 2350 This was with auto tuning. After disabling auto-bufpool tuning, I increased the bufpoolsize to 131072 (on a 2G system also running one other application) and notices the hit rate climb to 99.2% over the next two days. VMstat shows 'avm' of around 260000 pages, or just over a GIG. I therefore _assume_ the remainder constitutes buffer-cache. vmstat showed absolutely no paging for the period I monitored (about 6 hours during a working day), so would it be safe to further increase the bufpoolsize? Also, the customer noticed a TSM db backup which took 35 minutes two days ago, now completes in 12 minutes. Could that indicate a problem? TIA Dave