RE: TUX Crashes Under Stress

["Scottie M Brown" <scotbro us ibm com>]

Ingo,

I have been seeing a problem like this myself with 2.4.1-P6.  kswapd eats
up 100% cpu after about 12GB of data has been cached.  I plan to switch to
H3 today to see if the problem still exists.  Please let me know if there
is a fix for this problem.

Thanks,

Scott Brown - IBM PSG
Applications Performance Analysis
Phone: 919-543-5584
Fax: 919-254-9267



"Phil Oester" <phil@theoesters.com>@redhat.com on 02/28/2001 05:07:24 PM

Please respond to tux-list@redhat.com

Sent by:  tux-list-admin@redhat.com


To:   <tux-list@redhat.com>
cc:
Subject:  RE: TUX Crashes Under Stress



Ok - it was H2.  It wasn't a process eating up 100% cpu - it was the system
itself.  Top didn't show any process using a significant amount of cpu.

-----Original Message-----
From: tux-list-admin@redhat.com [mailto:tux-list-admin@redhat.com]On
Behalf Of Ingo Molnar
Sent: Wednesday, February 28, 2001 2:04 PM
To: tux-list@redhat.com
Subject: RE: TUX Crashes Under Stress



On Wed, 28 Feb 2001, Phil Oester wrote:

> In a possibly related situation, I'm using a Tux-patched 2.4.2 (forget
> the number, but its AC5 and was uploaded today) as a qmail server, and

(the latest, H3 patch is ac6-based.)

> had to rollback to a 2.2 kernel because after injecting a few thousand
> emails, *one* of the two processors gets pegged at 100% and brings the
> box to its knees. The box runs fine on 2.2.

hm, do you remember which process or processes used up 100% CPU time? Was
it kswapd? Or qmail?

     Ingo



_______________________________________________
tux-list mailing list
tux-list@redhat.com
https://listman.redhat.com/mailman/listinfo/tux-list




_______________________________________________
tux-list mailing list
tux-list@redhat.com
https://listman.redhat.com/mailman/listinfo/tux-list