[33529] in Hotline Meeting
mr-toad load average
daemon@ATHENA.MIT.EDU (Pat McCormick)
Thu May 16 00:01:12 1996
To: hotline@MIT.EDU
Cc: cprivate@MIT.EDU
Date: Thu, 16 May 1996 00:01:09 EDT
From: Pat McCormick <pmccormi@MIT.EDU>
Mr-toad has an abnormally high load average. Another consultant (shalabi)
noticed that it was running his animations slow today, and I just noticed
that load average with myself on console running one emacs, two xterms,
a clock and xload:
11:47pm up 2 days, 2:40, 2 users, load average: 1.35, 1.34, 1.33
I logged into air-traffic-controller, which has one user on console
and two people (myself and another) using the machine through telnet.
Load average is:
11:44pm up 1 day, 5:07, 3 users, load average: 0.48, 0.35, 0.22
On mr-toad, the only processes which seem to have a large CPU time are:
root 513 1 80 May 13 ? 589:05 /usr/Cadmin/bin/objectserver
root 520 1 0 May 13 ? 2:35 /usr/etc/mediad
On air-traffic-controller, these processes have times of 0:17 and 1:10,
respectively, so I think that they might be causing the increased load
on mr-toad.
The one difference between mr-toad and air-traffic-controller that
I know of is that mr-toad runs beta-sgisys and atc runs public-sgisys.
(Actually, these should be reversed, since 11-115 machines are supposed
to reflect cluster machines.)
Any ideas on this? I don't know what the above processes are supposed to
be doing, but I'm pretty certain that mr-toad is spending more time
than it should on them. Might this have something to do with the memory
errors we've been getting?
--Pat McCormick
