Re: Concurrent I/O (AIX 5.2)
- From: "Stuart Clowes" <stuart.clowes@xxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx
- Date: Thu, 29 Jun 2006 23:10:01 +0100
We have minperm=10%, maxperm=20%, maxclient=20%. Will look to reduce these
even further.
Our db_cache_size has increased from 1 GB to 2 GB, and we are still gently
increasing it.
We are, at the moment, not memory constrained, although I am starting to
chew up our spare memory by increasing db_cache_size. (use it or lose it....
;) )
agblksize=512 for the redo logs.
In fact, agblksize=512 for the database files as well - any comments on
whether this is 'wrong'?
Filesystems for redo and dbf are mounted cio. FILESYSTEM_IO_OPTIONS=SETALL.
Only thing we haven't done is implemented large memory pages. Do you know of
a metric that would reveal whether this could be beneficial for us?
Other related posts:
