[NTLUG:Discuss] Serious Problem

[Dan Carlson]

Here is a report from someone who shares your pain:

http://www.redhat.com/mailing-lists/ext3-users/msg03650.html

And the next message in the thread suggests the source of the problem and a
fix:

http://www.redhat.com/mailing-lists/ext3-users/msg03651.html

Another alternative might be to switch to the 2.4.19 kernel.  It has a very
large number of fixes, including a kjournald fix.

Dan Carlson

----- Original Message -----
From: "TJ Davis" <TJDavis at sagu.edu>
To: <discuss at ntlug.org>
Sent: Friday, August 16, 2002 11:55 AM
Subject: RE: [NTLUG:Discuss] Serious Problem


> Okay.  I guess I was a little too jumpy in sending the first message
because
> I found the error message in /var/log/messages.  Any ideas?  Here it is:
>
> Aug 16 11:21:51 webct2 kernel: ------------[ cut here ]------------
>
> Aug 16 11:21:51 webct2 kernel: kernel BUG at commit.c:535!
>
> Aug 16 11:21:51 webct2 kernel: invalid operand: 0000
>
> Aug 16 11:21:51 webct2 kernel: nls_iso8859-1 loop vfat fat ipchains
autofs
> eepro100 st usb-ohci usbcore ext3
>
> Aug 16 11:21:51 webct2 kernel: CPU: 0
>
> Aug 16 11:21:51 webct2 kernel: EIP: 0010:[<f88630e4>] Not tainted
>
> Aug 16 11:21:51 webct2 kernel: EFLAGS: 00010286
>
> Aug 16 11:21:51 webct2 kernel:
>
> Aug 16 11:21:51 webct2 kernel: EIP is at journal_commit_transaction [jbd]
> 0xb04 (2.4.18-3smp)
>
> Aug 16 11:21:51 webct2 kernel: eax: 0000001c ebx: 0000000a ecx: c02eee60
> edx: 00003a06
>
> Aug 16 11:21:51 webct2 kernel: esi: e8f46220 edi: ebd6a480 ebp: f6616000
> esp: f6617e78
>
> Aug 16 11:21:51 webct2 kernel: ds: 0018 es: 0018 ss: 0018
>
> Aug 16 11:21:51 webct2 kernel: Process kjournald (pid: 196,
> stackpage=f6617000)
>
> Aug 16 11:21:51 webct2 kernel: Stack: f8869eee 00000217 f64d6800 00000000
> 00000fd4 f6baf02c 00000000 f64ee1c0
>
> Aug 16 11:21:51 webct2 kernel: f54d01c0 000005c9 80000000 00000001
00000009
> f64d6964 f6d7b1e0 cc9f0de0
>
> Aug 16 11:21:51 webct2 kernel: f40ab420 f40b6ce0 f3f5ff20 f3fb93c0
f3fb9360
> f40b65c0 f40b6560 f01c0a20
>
> Aug 16 11:21:51 webct2 kernel: Call Trace: [<f8869eee>] .rodata.str1.1
[jbd]
> 0x26e
>
> Aug 16 11:21:51 webct2 kernel: [<c0124eb5>] update_process_times [kernel]
> 0x25
>
> Aug 16 11:21:51 webct2 kernel: [<c0116049>] smp_apic_timer_interrupt
> [kernel] 0xa9
>
> Aug 16 11:21:51 webct2 kernel: [<c010a77f>] do_IRQ [kernel] 0xdf
>
> Aug 16 11:21:51 webct2 kernel: [<c0119048>] schedule [kernel] 0x348
>
> Aug 16 11:21:51 webct2 kernel: [<f88657d6>] kjournald [jbd] 0x136
>
> Aug 16 11:21:51 webct2 kernel: [<f8865680>] commit_timeout [jbd] 0x0
>
> Aug 16 11:21:51 webct2 kernel: [<c0107286>] kernel_thread [kernel] 0x26
>
> Aug 16 11:21:51 webct2 kernel: [<f88656a0>] kjournald [jbd] 0x0
>
> Aug 16 11:21:51 webct2 kernel:
>
> Aug 16 11:21:51 webct2 kernel:
>
> Aug 16 11:21:51 webct2 kernel: Code: 0f 0b 5a 59 6a 04 8b 44 24 18 50 56
e8
> 4b f1 ff ff 8d 47 48
>
> T.J. Davis
> Southwestern A/G University
> Information Technology
> tjdavis at sagu.edu
> (972) 937-4010 ext. 1255
> 1 Timothy 4:12
>
> -----Original Message-----
> From: TJ Davis [mailto:TJDavis at sagu.edu]
> Sent: Friday, August 16, 2002 11:43 AM
> To: 'discuss at ntlug.org'
> Subject: [NTLUG:Discuss] Serious Problem
>
>
>
> Okay.  I have a serious problem going on with my RedHat Linux 7.3 server.
> This thing has been up for about 3 weeks now and today, for the second
time,
> the screen (tty1) filled up with error messages and it was still
basically
> running because I could get to the web page that it was serving but I
could
> not login to the console or remotely through SSH.  Samba died as well.  I
> also tried logging in through Webmin but it was dead as well.  My only
> option was to reboot and Lord knows I don't need that problem on my only
> Linux box. The error mentioned something about kjournald.  Unfortunately
I
> am still learning my way around log files and I have no idea as to where
I
> could find the logging of this error.  I guess for now if someone could
> point me to where to find the log of this error that would be great and
then
> I can go from there unless someone happens to know from the small details
> that I provided what the problem is.  Thank you.
>
> T.J. Davis
>
>