From: Giovanni Navarrette (gio@uslink.net)
Date: Wed May 28 2003 - 11:27:31 EDT
Hey everyone -
Does anyone know of a command or a way to trace why a process is chewing up
tons of disktime? Basically what I've got is a new server running
sendmail-8.12.9, and when its running and I'm using a spam program to send
messages to it, testing it for load I'm seeing 99% busy time on the disk,
and sendmail is slow as mud. This program sends 2 messages a second to
it..., so its putting it under decent stress, but the disk I/O is choking up
like mad. On this new system, this is what I see with iostat for that
drive...
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 139.7 0.0 968.0 0.0 1.8 0.0 12.8 0 99 c1t10d0s0
On our production system (older system), I see this:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.3 79.0 2.7 1051.5 0.0 1.3 0.0 16.0 0 48 c0t0d0s0
The production box is writing more to the drive..., but is less busy...,
does that mean that the new system just has a slower disk? I thought it was
a 10k RPM drive, but now that I think about it I haven't checked that yet
(note to self: check drive speed). When I shutdown sendmail, busy time goes
down to and stays at 0%. I've tried vmstat, top, ps, and iostat and I'm
having no luck narrowing down why the disk is so busy..., any ideas?
Thanks for ya time, have a g'day.
-
Giovanni Navarrette
USLink Internet Systems Administrator
e-Mail | gio@uslink.net
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:29 EDT