Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 3 Apr 2005 19:02:16 +0200
From:      Matthias Buelow <mkb@incubus.de>
To:        Emanuel Strobl <emanuel.strobl@gmx.net>
Cc:        freebsd-stable@freebsd.org
Subject:   Re: 5.4-prerelease - hanging under load
Message-ID:  <20050403170216.GA680@drjekyll.mkbuelow.net>
In-Reply-To: <200504031834.54592@harrymail>
References:  <424F5210.8060809@pldrouin.net> <20050403133200.GP87756@unixpages.org> <425014B2.9070208@pldrouin.net> <200504031834.54592@harrymail>

next in thread | previous in thread | raw e-mail | index | archive | help
Emanuel Strobl wrote:

>Huh? I thought by default it is "HIGH", but it would also explain the 
>experiences in the thread "cpufreq related RELENG_5 regression" 
>(http://www.freebsd.org/cgi/getmsg.cgi?fetch=389949+393881+/usr/local/www/db/text/2005/freebsd-stable/20050327.freebsd-stable)

But does that also affect desktop machines?  I'm not under the
impression that my CPU is running at a lower frequency.  The issue
I mentioned only seems to occur when doing larger block reads/writes
from/to disk (SATA), like untarring firefox.  I didn't notice it
so far when some kind of mixed disk access is going on, like with
find, compiling, etc.  Maybe that points to some locking issues in
the VM corner?  I mean, it's not dramatic, but still makes the
system appear a bit unpolished, when it happens.  In the past, with
pre-5.x releases, FreeBSD has been scheduling disk i/o and interactive
work smoothly, in my experience, so one didn't quite notice when
heavy disk access was going on, at least not if your interactive
programs weren't doing much disk i/o themselves.

mkb.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050403170216.GA680>