Re: ksoftirqd_cpu0 hogs CPU

# uname -a
Linux server01 2.4.25 #1 Wed Mar 3 00:01:28 AKST 2004 i686 unknown

how might I be able to tell you the "way it was built"? what
information is useful to you for this?

Dennis Byrne

----- Original Message -----
From: Adam Bultman <adamb@glaven.org>
Date: Monday, May 10, 2004 10:31 am
Subject: Re: ksoftirqd_cpu0 hogs CPU

> DENNIS BYRNE wrote:
>
> >Slackware 9.1
> >Problem #1: Every other hour ksoftirqd_cpu0 wakes up and
> consumes all
> >of the CPU cycles for 5 minutes. The machine stops serving web
> pages,
> >serving files, and sometimes my SSH connection is dropped.
> Google
> >searches don't turn up much here.
> >Problem #2: Apache 2.0.47 uses cronolog 1.6.2. When you restart
> >apache, it doesn't tell it's cronolog/process children to die.
> >Consequently, apache restarts, spawns new cronolog processes.
> There
> >are 22 idle cronolog processes running on this machine and it
> seems
> >that the only way to only solution is kill -9 or a restart.
> >
> >Does anyone know how Problem #1 could be related to problem #2?
> >
> >Dennis Byrne
> >
> >---------
> >To unsubscribe, send email to <aklug-request@aklug.org>
> >with 'unsubscribe' in the message body.
> >
> >
> >
> What kernel are you running?
>
> At work, we run all RedHat systems, and there's a bug report and a
> fix
> for this. Your kernel version, and the way it was built might be
> important in this case.
>
> Adam
>
>

---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Mon May 10 10:33:12 2004