Increase in buffer list

When the onstat -b command shows a list of buffers in addition to the summary, these lines represent buffers that are currently being modified. For several weeks we have experienced very few or no buffers on this list during routine production. This morning we are seeing anywhere from 200 to 300 buffers on this list and the number does not drop to zero after a checkpoint. There seems to be an unusual number of reads for the normal user sessions and some disks are up to 100% busy. We have made no changes to the DB. We have had satisfactory checkpoints, etc. with the present memory settings: SHMVIRTSIZE 1700000, BUFFERS 750000, LRUS 9, CLEANERS 9, LRU_MIN/MAX_DIRTY 1/3. What is the best approach to trouble-shoot this from the database standpoint? -- BillS Sent via Deja.com http://www.deja.com/ Before you buy.

onstat -u to find the session which is doing all the reads/writes onstat -g ses <session-id> + ask the user what they are doing. billybob wrote in message <8pqpod$slo$1@nnrp1.deja.com>... >When the onstat -b command shows a list of buffers in addition to the >summary, these lines represent buffers that are currently being >modified. For several weeks we have experienced very few or no buffers >on this list during routine production. This morning we are seeing >anywhere from 200 to 300 buffers on this list and the number does not >drop to zero after a checkpoint. There seems to be an unusual number >of reads for the normal user sessions and some disks are up to 100% >busy. We have made no changes to the DB. We have had satisfactory >checkpoints, etc. with the present memory settings: SHMVIRTSIZE >1700000, BUFFERS 750000, LRUS 9, CLEANERS 9, LRU_MIN/MAX_DIRTY 1/3. >What is the best approach to trouble-shoot this from the database >standpoint? > > >-- >BillS > > >Sent via Deja.com http://www.deja.com/ >Before you buy.