Here is a final piece of info on fsflush that I think is extremely
useful:
Checkout
http://www.Sun.COM:80/sunworldonline/swol-01-1996/swol-01-perf.htm
This is from Adrian Cockcroft's Book "Sun Performance and Tuning" from
SunPress:
autoup and tune_t_fsflushr
Unlike SunOS 4 when the update process does a full sync of memory to
disk every 30 seconds, Solaris 2 uses the fsflush daemon to spread
out the sync workload. Autoup is set to 30 seconds by default, amd
this is the maximum age of any memory resident filesystem pages that have
been modified. Unlike update, fsflush wakes up every 5 seconds (set by
tune_t_fsflushr) and checks a portion of memory on each invocation
(5/30 = one-sixth of total RAM by default). The pages are queued on the
same list that the pageout daemon uses and are formed into clustered
sequential writes.
On machines with more then a few hundred Mbytes of RAM, fsflush can take
over almost an entire CPU in the worts case, when very many pages are
being modified. This problem can be avoided by reducing the rate at which
fsflush checks memory. It should still always wake up every few seconds,
but autoup can be increased from 30 seconds to a few hundred seconds if
required. In most cases, the files that are being written are closed
before fsflush gets around to them. For NFS servers all writes are
synchronous so fsflush is hardly needed at all. Note the time and the CPU
usage of fsflush, then watch it later and see if its CPU usage is more
then five percent. If it is, increase autoup as showen.
set autoup=240 (in /etc/system)
-- http://www.acadiau.ca/cc/alan/
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:50 CDT