Lot of processes in "D" state with wchan = sync_buffer or sync_page
I have an 11203 RAC database and clusterware running on OEL 5.
One node is fine, the other node consistently crawls to a halt with these symptoms:
* top and vmstat show CPU in 90%+ wait (!) and high "b" queue
* ps shows a lot of processes in "D" state and the wchan for all of them is "sync_buffer" and/or "sync_page"
* this only seems to happen when I get EM DB Console up and running - and the db console doesn't run on the OK node.
* IOstat shows a lot of writes to the internal hard drive (All the datafiles are on external shared storage) but it doesn't look crazy high
One node is fine, the other node consistently crawls to a halt with these symptoms:
* top and vmstat show CPU in 90%+ wait (!) and high "b" queue
* ps shows a lot of processes in "D" state and the wchan for all of them is "sync_buffer" and/or "sync_page"
* this only seems to happen when I get EM DB Console up and running - and the db console doesn't run on the OK node.
* IOstat shows a lot of writes to the internal hard drive (All the datafiles are on external shared storage) but it doesn't look crazy high
0