D
dianne.evans@danwood.co.u
Guest
Following on from a colleague's recent question on our recurring database crashes due to Semaphores: I was wondering if people could assist with what a typical range of semaphore-wait should be, in "normal" behaviour. Our DB has been up for about a week now, and looking in PROMON R&D section, I can see the following values. 12/22/15 Activity: Other 11:16:21 12/15/15 12:37 to 12/22/15 11:12 (166 hrs 34 min) Total Per Min Per Sec Per Tx Commit 44492519 4452 74.20 1.00 Undo 247015 25 0.41 0.01 Wait on semaphore 376256 38 0.63 0.01 Non-blocking waits 25 0 0.00 0.00 Semaphore latch waits 43765 4 0.07 0.00 Flush master block 2097 0 0.00 0.00 Can anyone say whether the values of wait-on-semaphore = 38 per minute, or the semaphore-latch-waits of 4 per minute are OK, or extreme? The above is taken from our Live database, but for the last two days, both of our test databases (copies of live) have suffered similar semaphore problems. Any pointers would be a great help
Continue reading...
Continue reading...