Lost Files

Lost Files

[Alexander Waller]

Hi !
Yesterday it was the third time I lost important files on a
reiserfs-partition.
This time I had a reasonable backup.
So i could minimize the time my server was offline.
I use S.u.S.E 8.2.
The first to times reiserfsck was not able to repair any file. I found all
of them in the directory /lost+found, but that was not good for me. ( 11000
files there )
This time about 5000 files were in /lost+found and the rest in the right
place.

Is this a normal behavior or do you think I have a hardware problem ?
Is the S.u.S.E implementation wrong ?
Shall I change to ext3 ? ( It´s slower, I know )

Whats the right strategie to recover a defetive partition ?

Alex.


-- 

WebPoint Internet Solutions
Alexander Waller
Vorarlberger Wirtschaftspark
A-6840 Götzis
 
www.webpoint.at
tel.  +43 5523 58244
fax.  +43 5523 58255
mobil +43 676 4121128
 
webmaster-team  : webmaster@webpoint.at
hostmaster-team : hostmaster@webpoint.at

***                                     ***
*       WebPoint Internet Solutions       *
***                                     ***
 

Re: Lost Files

[Hans Reiser]

Alexander Waller wrote:

>Hi !
>Yesterday it was the third time I lost important files on a
>reiserfs-partition.
>This time I had a reasonable backup.
>So i could minimize the time my server was offline.
>I use S.u.S.E 8.2.
>The first to times reiserfsck was not able to repair any file. I found all
>of them in the directory /lost+found, but that was not good for me. ( 11000
>files there )
>This time about 5000 files were in /lost+found and the rest in the right
>place.
>
>Is this a normal behavior or do you think I have a hardware problem ?
>Is the S.u.S.E implementation wrong ?
>Shall I change to ext3 ? ( It´s slower, I know )
>
>Whats the right strategie to recover a defetive partition ?
>
>Alex.
>
>
>  
>
It is not normal.  It is very likely to be hardware.

-- 
Hans

Re: Lost Files

[Carl-Daniel Hailfinger]

Hans Reiser wrote:
> Alexander Waller wrote:
> 
>> Hi !
>> Yesterday it was the third time I lost important files on a
>> reiserfs-partition.
>> This time I had a reasonable backup.
>> So i could minimize the time my server was offline.
>> I use S.u.S.E 8.2.
>> The first to times reiserfsck was not able to repair any file. I found
>> all
>> of them in the directory /lost+found, but that was not good for me. (
>> 11000
>> files there )
>> This time about 5000 files were in /lost+found and the rest in the right
>> place.
>>
>> Is this a normal behavior or do you think I have a hardware problem ?
>> Is the S.u.S.E implementation wrong ?
>> Shall I change to ext3 ? ( It´s slower, I know )
>>
>> Whats the right strategie to recover a defetive partition ?
>>
>> Alex.
>>
>>
>>  
>>
> It is not normal.  It is very likely to be hardware.

Alex: Can you check /var/log/messages if there is any line containing
"error", "oops" or something like that?


Carl-Daniel
-- 
http://www.hailfinger.org/

Re: Lost Files

[Christian Mayrhuber]

Alexander Waller wrote:
> Hi !
> Yesterday it was the third time I lost important files on a
> reiserfs-partition.
> This time I had a reasonable backup.
> So i could minimize the time my server was offline.
> I use S.u.S.E 8.2.
> The first to times reiserfsck was not able to repair any file. I found all
> of them in the directory /lost+found, but that was not good for me. ( 11000
> files there )
> This time about 5000 files were in /lost+found and the rest in the right
> place.
> 
> Is this a normal behavior or do you think I have a hardware problem ?
> Is the S.u.S.E implementation wrong ?
> Shall I change to ext3 ? ( It´s slower, I know )
> 
> Whats the right strategie to recover a defetive partition ?
> 
> Alex.
> 
> 

I once had a similiar problem, the problem was the AMI Megaraid (now LSI 
   Logic) controller. It didn't detect failed disks which resulted in 3 
failed disks of 5 in a raid 5 array and an unused hot spare.
I had to build a completely new raid array. The thrust in megaraid is gone.

If you want to try various methos for disk recovery, you will have to
use dd_rescue to do a complete backup of the data on your disk into a file.
Use the newest reiserfsck.
Copy the backup to a new name, set up a loop back device and run 
reiserfsck --check on it. Then do what reiserfsck suggests.
If --rebuild-tree recovers to less files you can try --scan-whole-partition.
I don't know a way how to recover the filenames, though.
On a mailstore the files residing in lost+found could be scanned by a 
script to be placed into the right mailboxes, but that doesn't work
on a webspace :-(

-- 
lg, Chris

Re: Lost Files

[Robert_Thilsted]

I have 40+ servers with 18 to 500Gig partitions, hardware and software RAID
1 and 5, Pentium III - Pentium 4 - AMD Opteron, 1 to 4 processors, and all
running SuSE 8.2 (moving to 9.0) using reiserfs.  I have never had a case
of data corruption or loss in the last two years.  Your problem could be
hardware but the first thing I would do is to try SuSE 9.0.  From my
experience, definitely not normal.



|---------+---------------------------->
|         |           "Alexander       |
|         |           Waller"          |
|         |           <a.waller@webpoin|
|         |           t.at>            |
|         |                            |
|         |           10/29/03 01:06 AM|
|         |                            |
|---------+---------------------------->
  >---------------------------------------------------------------------------------------------------------------|
  |                                                                                                               |
  |       To:       <reiserfs-list@namesys.com>                                                                   |
  |       cc:       <webmaster@webpoint.at>                                                                       |
  |       Subject:  Lost Files                                                                                    |
  >---------------------------------------------------------------------------------------------------------------|




Hi !
Yesterday it was the third time I lost important files on a
reiserfs-partition.
This time I had a reasonable backup.
So i could minimize the time my server was offline.
I use S.u.S.E 8.2.
The first to times reiserfsck was not able to repair any file. I found all
of them in the directory /lost+found, but that was not good for me. ( 11000
files there )
This time about 5000 files were in /lost+found and the rest in the right
place.

Is this a normal behavior or do you think I have a hardware problem ?
Is the S.u.S.E implementation wrong ?
Shall I change to ext3 ? ( It´s slower, I know )

Whats the right strategie to recover a defetive partition ?

Alex.


--

WebPoint Internet Solutions
Alexander Waller
Vorarlberger Wirtschaftspark
A-6840 Götzis

www.webpoint.at
tel.  +43 5523 58244
fax.  +43 5523 58255
mobil +43 676 4121128

webmaster-team  : webmaster@webpoint.at
hostmaster-team : hostmaster@webpoint.at

***                                     ***
*       WebPoint Internet Solutions       *
***                                     ***

reiserfs error messages

[Vincent Ndi]

Hi everyone,

My system is keeps reporting some reiserfs errors messages I don't really
understand.Can someone advice me about the following error messages (how to
interpret/correct them)?

1) kernel: vs-13048: reiserfs_iget: key in inode [2 570 0 0] and key in
entry [39 570 0 0] do not match

2)kernel: vs-6020: check_leaf_block_head: invalid free space level=1,
nr_items=21, free_space=65112 rdkey [5654 5656 2511 -1]<0>Kernel panic:
REISERFS: panic (device sb == 0): vs-6020: check_leaf_block_head: invalid
free space level=1, nr_items=21, free_spa

any clue will be helpful


regards

Vincent

Re: reiserfs error messages

[Vitaly Fertman]

On Friday 31 October 2003 12:02, Vincent Ndi wrote:
> Hi everyone,
>
> My system is keeps reporting some reiserfs errors messages I don't really
> understand.Can someone advice me about the following error messages (how to
> interpret/correct them)?
>
> 1) kernel: vs-13048: reiserfs_iget: key in inode [2 570 0 0] and key in
> entry [39 570 0 0] do not match
>
> 2)kernel: vs-6020: check_leaf_block_head: invalid free space level=1,
> nr_items=21, free_space=65112 rdkey [5654 5656 2511 -1]<0>Kernel panic:
> REISERFS: panic (device sb == 0): vs-6020: check_leaf_block_head: invalid
> free space level=1, nr_items=21, free_spa
>
> any clue will be helpful

Would you run reiserfsck on the problem partition? It seems it has some 
corruptions. Use the latest reisrfsprogs:
ftp.namesys.com/pub/reiserfsprogs/reiserfsprogs-3.6.11.tar.gz

-- 
Thanks,
Vitaly Fertman

RE: reiserfs error messages

[Vincent Ndi]

Hi,

Ok I plan to run reiserfsck but on these errors messages, there's no
indication about the corrupted partition.Can you please explain what these
messages mean.

regards
Vincent

-----Original Message-----
From: Vitaly Fertman [mailto:vitaly@namesys.com]
Sent: Friday, October 31, 2003 10:08 AM
To: Vincent Ndi
Cc: reiserfs-list@namesys.com
Subject: Re: reiserfs error messages


On Friday 31 October 2003 12:02, Vincent Ndi wrote:
> Hi everyone,
>
> My system is keeps reporting some reiserfs errors messages I don't really
> understand.Can someone advice me about the following error messages (how
to
> interpret/correct them)?
>
> 1) kernel: vs-13048: reiserfs_iget: key in inode [2 570 0 0] and key in
> entry [39 570 0 0] do not match
>
> 2)kernel: vs-6020: check_leaf_block_head: invalid free space level=1,
> nr_items=21, free_space=65112 rdkey [5654 5656 2511 -1]<0>Kernel panic:
> REISERFS: panic (device sb == 0): vs-6020: check_leaf_block_head: invalid
> free space level=1, nr_items=21, free_spa
>
> any clue will be helpful

Would you run reiserfsck on the problem partition? It seems it has some
corruptions. Use the latest reisrfsprogs:
ftp.namesys.com/pub/reiserfsprogs/reiserfsprogs-3.6.11.tar.gz

--
Thanks,
Vitaly Fertman

HELP on reiserfs

[Vincent Ndi]

 Hi everyone,

My system is keeps reporting some reiserfs errors messages I don't really
 understand.Can someone advice me about the following error messages (how to
 interpret/correct them)?

 1) kernel: vs-13048: reiserfs_iget: key in inode [2 570 0 0] and key in
 entry [39 570 0 0] do not match

 2)kernel: vs-6020: check_leaf_block_head: invalid free space level=1,
 nr_items=21, free_space=65112 rdkey [5654 5656 2511 -1]<0>Kernel panic:
 REISERFS: panic (device sb == 0): vs-6020: check_leaf_block_head: invalid
 free space level=1, nr_items=21, free_spa

 any clue will be helpful.

Regards
Vincent

Re: reiserfs error messages

[Vitaly Fertman]

On Friday 31 October 2003 12:16, Vincent Ndi wrote:
> Hi,
>
> Ok I plan to run reiserfsck but on these errors messages, there's no
> indication about the corrupted partition.Can you please explain what these
> messages mean.
>
> regards
> Vincent
>
> On Friday 31 October 2003 12:02, Vincent Ndi wrote:
> > Hi everyone,
> >
> > My system is keeps reporting some reiserfs errors messages I don't really
> > understand.Can someone advice me about the following error messages (how
>
> to
>
> > interpret/correct them)?
> >
> > 1) kernel: vs-13048: reiserfs_iget: key in inode [2 570 0 0] and key in
> > entry [39 570 0 0] do not match

this is probably nothing to worry about if you use nfs, if not, this is a
corruption in the directory entry pointing to the file (inode).

This message was removed in later kernel versions.

> > 2)kernel: vs-6020: check_leaf_block_head: invalid free space level=1,
> > nr_items=21, free_space=65112 rdkey [5654 5656 2511 -1]<0>Kernel panic:
> > REISERFS: panic (device sb == 0): vs-6020: check_leaf_block_head: invalid
> > free space level=1, nr_items=21, free_spa

a corruption in some formatted reiserfs block, wrong free space in the 
block, probably some other corruptions exist also.

-- 
Thanks,
Vitaly Fertman