RE: v$sql 'db block changes'

> in terms of cpu, elapsed_time, buffer_gets and disk_reads are all
SELECT statements,
> so I know they're not causing the write activity.

Brandon,

Some SELECT statements do require writing to TEMP (due to SORT or HASH operations). And since direct writes to TEMP do not have to be (redo)logged, they are done fairly rapidly which can result in the kind of problems with RAID 5 that you are seeing.... Your best option is to track progress of V$SESSTAT counters such as 'physical writes direct'. If you like using the Wait interface, then you could look at sessions with high waits on 'direct path write' (in 10gr2 this is split out as 'direct path write temp').

See the earlier thread 'Simulate AWR like repository in 9i' for some additional thoughts on this.

Cheers,
John Kanagaraj <><
DB Soft Inc
Phone: 408-970-7002 (W)  

Co-Author: Oracle Database 10g Insider Solutions http://www.samspublishing.com/title/0672327910  

• The opinions and facts contained in this message are entirely mine and do not reflect those of my employer or customers ** -- http://www.freelists.org/webpage/oracle-l