New error when running script on ReiserFS partition

New error when running script on ReiserFS partition

[Alexandre Ratti]

Hello,


I got a new error when my backup script ran today. (See [1] for previous 
message sent on Tuesday). This script runs on a large ReiserFS partition. 
System is Debian Woody with recompiled 2.4.19 kernel. System did not crash.

[1] http://marc.theaimsgroup.com/?t=103955025600002&r=1&w=2

I saw with ps that the kupdated process was dead (defunct).

I'm not sure this is ReiserFS-related, though these errors only seem to 
happen when the script runs. Can someone confirm, based on this syslog snippet?


Thanks.

Alexandre

----------
Dec 12 13:10:01 p120 /USR/SBIN/CRON[1612]: (backup) CMD 
(/usr/local/bin/roller.py)
Dec 12 13:15:01 p120 kernel: Unable to handle kernel paging request at 
virtual address 0002dd4c
Dec 12 13:15:01 p120 kernel:  printing eip:
Dec 12 13:15:01 p120 kernel: c0186f77
Dec 12 13:15:01 p120 kernel: *pde = 00000000
Dec 12 13:15:01 p120 kernel: Oops: 0000
Dec 12 13:15:01 p120 kernel: CPU:    0
Dec 12 13:15:02 p120 kernel: EIP:    0010:[can_dirty+87/144]    Not tainted
Dec 12 13:15:02 p120 kernel: EFLAGS: 00010206
Dec 12 13:15:02 p120 kernel: eax: 0002dd48   ebx: c383c75c   ecx: 
00000001   edx: c38713ec
Dec 12 13:15:02 p120 kernel: esi: 0001fe7e   edi: 00001601   ebp: 
c38193bc   esp: c10e9f3c
Dec 12 13:15:02 p120 kernel: ds: 0018   es: 0018   ss: 0018
Dec 12 13:15:02 p120 kernel: Process kupdated (pid: 7, stackpage=c10e9000)
Dec 12 13:15:02 p120 kernel: Stack: c383c75c 00000004 00000000 c0184f29 
c383c75c
  c38193bc 00000004 00000005
Dec 12 13:15:02 p120 kernel:        00000000 00000001 c38193fc 00000000 
00000000 c01850da c2c89000 c38193bc
Dec 12 13:15:02 p120 kernel:        c10e9fa4 c2c89000 00000004 3df87dc2 
c0187178 c2c89000 c2c89000 c2c89044
Dec 12 13:15:02 p120 kernel: Call Trace: [kupdate_one_transaction+177/508] 
[reiserfs_journal_kupdate+102/148] [flush_old_commits+320/336] 
[reiserfs_write_sup
er+21/32] [sync_supers+188/240]
Dec 12 13:15:02 p120 kernel:   [sync_old_buffers+12/68] [kupdate+217/252] 
[kernel_thread+40/56]
Dec 12 13:15:02 p120 kernel:
Dec 12 13:15:02 p120 kernel: Code: 83 78 04 00 74 1f 8b 40 0c 85 c0 7e 18 
83 3a 00 74 13 8b 42
----------


>Date: Wed, 11 Dec 2002 11:28:25 +0100
>To: Oleg Drokin <green@namesys.com>
>From: Alexandre Ratti <alex@gabuzomeu.net>
>Subject: Re: "get_cnode failed" error
>Cc: reiserfs-list@namesys.com, mason@namesys.com

>At 10:26 11/12/2002 +0300, Oleg Drokin wrote:
>>On Tue, Dec 10, 2002 at 08:55:08PM +0100, Alexandre Ratti wrote:
>> > The FS looks OK, but I'm curious. What does this error mean? It is 
>> serious?
>>
>>This error means there was some journaling code screwup. So FS panicked 
>>before it could do any harm to on-disk data.
>>What does your script does with cp? Do you have at least some aproximate 
>>ideas on what it copied and from where to where?
>
>Yes, this is a backup script in Python. It first duplicate an existing 
>directory tree into a new tree using hard links, then it synchronise the 
>new tree with files located on another server using rsync. This create an 
>incremental backup. See [1] for theory and [2] for code.
>
>[1] http://www.mikerubel.org/computers/rsync_snapshots/
>[2] http://web2.airmail.net/lelie/roller.py
>
>I think the script was running "cp -al old_dir new_dir" to duplicate a 
>directory tree (16 GB) on the same partition. This is a large (120 GB) 
>partition I use to back up files; it is 30% full.
>
>The script ran 99 times without any glitch. Yesterday, after the error, I 
>upgraded the system to 2.4.19 and I ran the script again; it ran with no 
>error.

Re: New error when running script on ReiserFS partition

[Oleg Drokin]

Hello!

On Thu, Dec 12, 2002 at 02:54:33PM +0100, Alexandre Ratti wrote:

> I got a new error when my backup script ran today. (See [1] for previous 
> message sent on Tuesday). This script runs on a large ReiserFS partition. 
> System is Debian Woody with recompiled 2.4.19 kernel. System did not crash.
> I saw with ps that the kupdated process was dead (defunct).
> I'm not sure this is ReiserFS-related, though these errors only seem to 
> happen when the script runs. Can someone confirm, based on this syslog 
> snippet?

While this is certainly a piece of reiserfs code, the crash address makes no
sence at all.
Can you please run memtest86 for some time to confirm that your memory is still
ok?

Bye,
    Oleg

Re: New error when running script on ReiserFS partition

[Alexandre Ratti]

Hello Oleg,


At 17:07 12/12/2002 +0300, Oleg Drokin wrote:
>On Thu, Dec 12, 2002 at 02:54:33PM +0100, Alexandre Ratti wrote:
>
> > I got a new error when my backup script ran today. (See [1] for previous
> > message sent on Tuesday). This script runs on a large ReiserFS partition.
> > System is Debian Woody with recompiled 2.4.19 kernel. System did not crash.
> > I saw with ps that the kupdated process was dead (defunct).
> > I'm not sure this is ReiserFS-related, though these errors only seem to
> > happen when the script runs. Can someone confirm, based on this syslog
> > snippet?
>
>While this is certainly a piece of reiserfs code, the crash address makes no
>sence at all.
>Can you please run memtest86 for some time to confirm that your memory is 
>still ok?

It's running now; I'll let you know what the results are. (I last ran 
memtest86 in October when I set up this box and the memory was fine at that 
time.)


Cheers

Alexandre

Re: New error when running script on ReiserFS partition

[Alexandre Ratti]

Hello,


>Date: Thu, 12 Dec 2002 15:27:07 +0100
>From: Alexandre Ratti <alex@gabuzomeu.net>

>At 17:07 12/12/2002 +0300, Oleg Drokin wrote:
>>On Thu, Dec 12, 2002 at 02:54:33PM +0100, Alexandre Ratti wrote:
>> > I got a new error when my backup script ran today. (See [1] for previous
>> > message sent on Tuesday). This script runs on a large ReiserFS partition.
>> > System is Debian Woody with recompiled 2.4.19 kernel.

>>While this is certainly a piece of reiserfs code, the crash address makes no
>>sence at all. Can you please run memtest86 for some time to confirm that 
>>your memory is still ok?
>
>It's running now; I'll let you know what the results are. (I last ran 
>memtest86 in October when I set up this box and the memory was fine at 
>that time.)

I ran tests and watched the system for some time. The memory was fine, but 
I believe now that the problem was due to an old, flaky ISA graphic card.

The backup script has been running fine for more than a week now, so the 
problem looks fixed.


Thanks for your input.

Alexandre


[1] History: http://marc.theaimsgroup.com/?t=103955025600002&r=1&w=2