My original posting:
On our Sparc 1000 running Solaris 2.3, patch 101318-70, and
Oracle 7.0.16 under a heavy load, 'sar -b 30 1000' reports
%wcache being zero most of the time whereas %rcache almost
always 100. Is it abnormal or does it simply need tuning of
such parameters as nhbuf, p_nbuf, and bufhwm?
Sincere thanks to michael@ide.com (Michael K. Glass) who suggested
monitoring wcache with the DB server not running, and to
Glenn.Satchell@uniq.com.au (Glenn Satchell - Uniq Professional Services)
who was so kind to inquire additional info and recommend
You might like to look at Prestoserve, which provides a non-volatile
memory buffer for the disks and can dramatically improve write
performance. Alternatively look at Solstice Disksuite 4.0 and the
logging file system. Neither of these will work with raw partitions
though.
Also I only accelerated the partitions where the databases were. The
default is to accelerate all UFS filesystems, but this wastes cache
space. I had a couple of "hot" filesystems where I wanted the
acceleration most. I especially don't think you need to presto /,
/usr, /var or /opt since there is usually very little write activity
on these anyway.
The hardware has been ordered and shall be installed soon.
Anchi
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:32 CDT