2.4.18-pre3 ReiserFS crashes not for the first time...

DISCUSSION Grzegorz Prokopski 1 participant 2 messages Jan 20, 2002

Share & tools

Grzegorz Prokopski

author

2002-01-20 18:08 UTC #0 lore raw

Hi!

First I must say I have not sent anythink like that to LKML yet, but gonna do my best to give as much and as good information as I can.

Background:
I am using Debian unstable/SID. Because of some problems it sometimes shutsdown uncleanly - after logiing out from GNOME computer just freezes. It is unstable, normally I am using NVidia binary driver and VMware binary kernel modules so I wasn't surprised by this.

But I also use ReiserFS on LVM (only root is outside LVM), so I _was_ surprised when filesystem related problems came up.

It was so bad I had to use fsck.reiserfs --rebuild-tree on unmounted partitions in single user mode.

Problem:
Today I was doing dist-upgrade - it froze my computer. I *disabled loading all binary modules*, restarted my machine and run the process again. Below You can see what it caused:

Configuring bison (1.31-2)...

vs-7050: new entry is found, new inode == 0
invalid operand: 0000
CPU:    0
EIP:    0010:[<d081c239>]             Not tainted
EFLAGS: 00010286
    eax: 0000002f  ebx: d082bea0  ecx: ffffe061  edx: c4aea000
    esi: sffe8000  edi: ffffffef  ebp: c4aebf0c  esp: c4aebcec
    ds: 0018  es: 0018  ss: 0018
Process update-alternat (pid: 542, stacpage = c4aeb000)
Stack: d082d79a d0831b20 d082bea0 c4aebd10 c4a129a0 caebdd8 d0812299

cffe8000
d082bea0 c4a12a1c c4a12a1c c4a129a0 c49ac060 00000000 00000000 00000000 00000000 00000000 00000000 00000000 d08301cf 00000000 00000039 00013850
Calltrace: [<d082d79a>] [<d0831b20>] [<d082bea0>] [<d0812299>] [<d082bea0>]
[<d08301cf>] [<d0828baf>] [<c013cd99>] [<c013cf96>] [<c0106f43>]

Code: 0f 0b 68 20 1b 83 d0 85 f6 74 0f 0f b7 46 08 50 e8 ba 4d 91

/var/lib/dpkg/info/bison.postinst: line 10: 542 Segmentation fault

the 10th line of the script is just "fi"

Other stuff:
I am also getting strange errors like denying access to some files while doing ls on a directory content - as root of course.

System data:
kernel 2.4.18-pre3 with latest LVM patch and latest kernel preemtion patch (as of 15.I.2002)

Own thoughts:
I think that's all. Just wanna say I a bit dissapointed with reiserfs. When installing it (took me 2 days to do the transition from lilo, ext2 to grub, LVM and ReiserFS ;-) I thought it is meant to be immune against system crashes and power failures.

Now gotta go into single user mode now and --rebuild-tree again on every each partition... Maybe anyone has a better idea ?

I will be glad to see any responses, please Cc: them to me if You please ( mailto:[email protected] ) - or give me an address of any _live_ LKML news gateway.

Best regards
Grzegorz Prokopski

Grzegorz Prokopski

author

2002-01-20 18:59 UTC #1 lore raw

W liście z nie, 20-01-2002, godz. 17:09, Grzegorz Prokopski pisze:

Quoted 5 lines

Hi!
First I must say I have not sent anythink like that to LKML yet, but gonna do my best to give as much and as good information as I can.

Thanks to Mark Hahn for pointing me out that I should use ksymoops (I was thinking about it, but didn't really knew what to do with it)

Here it goes:

greg:~# ksymoops ./reiserfs.oops 
ksymoops 2.4.3 on i586 2.4.18-p3lvmpe.  Options used
     -V (default)
     -k /proc/ksyms (default)
     -l /proc/modules (default)
     -o /lib/modules/2.4.18-p3lvmpe/ (default)
     -m /boot/System.map-2.4.18-p3lvmpe (default)

Quoted 1 line

Quoted 1 line
Quoted 1 line
Quoted 1 line
those are right for my system I think

Warning (expand_objects): object /lib/modules/2.4.18-p3lvmpe/kernel/fs/reiserfs/reiserfs.o for module reiserfs has changed since load

Quoted 3 lines

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
this is the right module IMO - I use cramfs with modules to boot the machine - above modules dir was used to create the image from which reiserfs module was loaded

invalid operand: 0000
CPU:    0
EIP:    0010:[<d081c239>]             Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010286
 eax: 0000002f  ebx: d082bea0  ecx: ffffe061  edx: c4aea000
 esi: sffe8000  edi: ffffffef  ebp: c4aebf0c  esp: c4aebcec
 ds: 0018  es: 0018  ss: 0018
 Stack: d082d79a d0831b20 d082bea0 c4aebd10 c4a129a0 caebdd8 d0812299
 cffe8000
 d082bea0 c4a12a1c c4a12a1c c4a129a0 c49ac060 00000000 00000000 00000000
 00000000 00000000 00000000 00000000 d08301cf 00000000 00000039 00013850
  [<d08301cf>] [<d0828baf>] [<c013cd99>] [<c013cf96>] [<c0106f43>]
 Code: 0f 0b 68 20 1b 83 d0 85 f6 74 0f 0f b7 46 08 50 e8 ba 4d 91

Quoted 1 line

Quoted 1 line
EIP; d081c238 <[reiserfs]reiserfs_panic+28/54> <=====

Code; d081c238 <[reiserfs]reiserfs_panic+28/54>
00000000 <_EIP>:
Code; d081c238 <[reiserfs]reiserfs_panic+28/54> <=====
0: 0f 0b ud2a <=====
Code; d081c23a <[reiserfs]reiserfs_panic+2a/54>
2: 68 20 1b 83 d0 push $0xd0831b20
Code; d081c23e <[reiserfs]reiserfs_panic+2e/54>
7: 85 f6 test %esi,%esi
Code; d081c240 <[reiserfs]reiserfs_panic+30/54>
9: 74 0f je 1a <_EIP+0x1a> d081c252
<[reiserfs]reiserfs_panic+42/54>
Code; d081c242 <[reiserfs]reiserfs_panic+32/54>
b: 0f b7 46 08 movzwl 0x8(%esi),%eax
Code; d081c246 <[reiserfs]reiserfs_panic+36/54>
f: 50 push %eax
Code; d081c248 <[reiserfs]reiserfs_panic+38/54>
10: e8 ba 4d 91 00 call 914dcf <_EIP+0x914dcf> d1131006
<END_OF_CODE+15c928/????>

2 warnings issued. Results may not be reliable.
greg:~#

HTH
Grzegorz Prokopski