xfs
[Top] [All Lists]

Re: Kernel oops on new mailserver

To: Paul Schutte <paul@xxxxxxxxxxx>
Subject: Re: Kernel oops on new mailserver
From: John Wingenbach <john@xxxxxxxxxxxxxx>
Date: Mon, 18 Mar 2002 16:57:27 -0500
Cc: XFS mailing list <linux-xfs@xxxxxxxxxxx>
References: <3C965130.2CAB0A19@it.up.ac.za>
Sender: owner-linux-xfs@xxxxxxxxxxx
I couldn't bet the farm on it. But every time I see someone using the stock
RedHat egcs compiler I cringe.  egcs-2.91.66 from Redhat has caused more
problems due to an optimizer bug.  There are several news groups where it is
the root cause of problems in several apps.  Upgrade your compiler (and
rebuild your kernel) and that very well may eliminate your problems.

-- John Wingenbach

Paul Schutte wrote:

> Hi,
>
> I have set up a new mailserver, consisting of the following:
>
> Dell PE2550
> 1G RAM
> 2 x 1.133GHz CPUs
> 4x18Gb Seagate cheeta's in RAID 10
>
> kernel 2.4.18 from cvs checked out on March 14 2002.
> gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release)
> ksymoops 2.4.3
> Debian 3.0 (woody)
> JFS 1.0.15 were also patched into the kernel, but no jfs partitions were
> mounted at the time.
> I also compiled a kernel without any other filesystem in it. Not even
> ext2, but also got the problem.
> I did'nt had a terminal connected at that time, so I don't have an oops
> where XFS was the only
> filesystem.
>
> The kernel oopsed several times lately.
> I caught an oops, but couldn't do stack traces.
> I caught another one, this time with traces on both CPUs.
> The e100 ethernet module from intel was loaded in the first oops.
> I reverted back to the stock eepro100 module in the hope that it might
> solve the problem.
> The second oops was with the eepro100 module loaded
> No other modules were loaded.
>
>

Kernel trace removed...

>
>
> Pleases let me know if there is anything else that I must check.
>
> Paul Schutte


<Prev in Thread] Current Thread [Next in Thread>