> How are "free buffer waits" (below) to be addressed? Your DBWR probably isn't able to write dirty buffers to disk fast enough. See if you have disk_asynch_io = true set. If you're running on regular cooked filesystem (not QuickIO) then you should probably increase the number of db writer processes or dbwr slaves (instead of using OS simulated asynch IO). Btw, if you are on raw devices or QuickIO and using kernelized asynch IO, I would still keep the number of dbwr's greater than 1 in heavily IO intensive environment. Also, number of buffer cache LRU latches (controlled by _db_block_lru_latches) might become a bottleneck in your env (lots of CPUs + heavy IO). Also, you might want to increase your buffer cache (tm ;) to give dbwr a little more breathing room. Btw, why do you want to have separate 4 controllers for log files? I get the need for 2 ones (the IO generated by archiver won't disturb lgwr's IO), but why the additional 2? Just for having more time for archiver before lgwr hits the same controller again or are you using software striping over different controllers (or multiplexing) for redologs? Tanel. ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------