Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 7 Sep 2000 16:45:31 -0400 (EDT)
From:      Vivek Khera <khera@kciLink.com>
To:        stable@FreeBSD.ORG
Subject:   Re: Odd Load average problem
Message-ID:  <14775.65131.298732.839601@onceler.kciLink.com>
In-Reply-To: <Pine.BSF.3.96.1000908063510.20849B-100000@gaia.nimnet.asn.au>
References:  <39B7F42D.7BA8BDB9@columbus.rr.com> <Pine.BSF.3.96.1000908063510.20849B-100000@gaia.nimnet.asn.au>

next in thread | previous in thread | raw e-mail | index | archive | help
>>>>> "IS" == Ian Smith <smithi@nimnet.asn.au> writes:

IS> For what it's worth, I've noticed this occasionally over a couple of
IS> years on a 2.2.6 box; same, seems like 1.00 has been added to the load
IS> average, lasts for hours, goes away.  Thought it'd be fixed ages ago :)

There was a big discussion about this on the bsdi-users mailing list a
year or so ago.  The gist of the problem was that on x86 architecture,
you don't have enough timers to measure the system without affecting
the system.  Something about needing to make the resolution of the
measuring timer be relatively prime to the scheduling timer.  Thus, it
is possible for the two logical timers to get in sync and artifically
increase the load.  Usually running some CPU intensive job breaks the
synchronization.

This is from memory, so may be slightly inaccurate, and may not apply
to FreeBSD, though I suspect it does considering the issue was with
the hardware capabilities.

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Vivek Khera, Ph.D.                Khera Communications, Inc.
Internet: khera@kciLink.com       Rockville, MD       +1-301-545-6996
GPG & MIME spoken here            http://www.khera.org/~vivek/


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14775.65131.298732.839601>