Linux 2.4 -> 2.6 migration

Linux 2.4 -> 2.6 migration

[Sebastian Kaps]

Hi!

Is there something concerning ReiserFS I should know when migrating from
Linux 2.4 to Linux 2.6?

I'm asking because a few days ago I downloaded and compiled 2.6.0-test9
just for curiosity. It booted fine and I went back to my "standard"
kernel, which was 2.4.23-pre5 at that time. Right after booting the
2.4.x kernel, I got lots of the following messages in my logs:
,----
| kernel: ide2(33,3):vs-4080: reiserfs_free_block: free_block (2103:463529)[dev:blocknr]: bit already cleared
`----
The block number always differed. Gladly, "reiserfsck --rebuild-tree"
helped to cure that, but there were lots of unusable files after that
(e.g. parts of files overwritten with parts of other files).

Is this suppose to happen? What caused this error?

Another thing: I had to "hard reboot" my system a few times in the last
few months. I also experienced that sometimes parts of files were
overwritten with 0x00s or parts of other files. Can someone explain to
me how this can happen?

-- 
Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Eric Whiting]

I was running 2.6.0test6 on my desktop. 

I installed suse9 and ran their 2.4.21-99 kernel and had a lot of reiserfs
corruption on a existing partition. I also had to --rebuild-tree on that
parition. I'm runnning 2.6.0-test9 right now.

2.6->2.4 was not a happy thing for my box..

I thought it was something in how I did the upgrade. But maybe something else is
happening here... Somewhere in my logs I see this:

eric




Nov  4 15:32:13 windriver kernel: vs-6030: check_internal_block_head: invalid
item number level=2, nr_items=
170, free_space=65520 rdkey kernel BUG at prints.c:334!
Nov  4 15:32:13 windriver kernel: invalid operand: 0000 2.4.21-99-athlon #1 Wed
Sep 24 13:34:32 UTC 2003
Nov  4 15:32:13 windriver kernel: CPU:    0
Nov  4 15:32:13 windriver kernel: EIP:   
0010:[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers
/s+4275657781/96]    Not tainted
Nov  4 15:32:13 windriver kernel: EIP:    0010:[<c1995c35>]    Not tainted
Nov  4 15:32:13 windriver kernel: EFLAGS: 00010292
Nov  4 15:32:13 windriver kernel: eax: 00000069   ebx: 00000000   ecx:
c037e4d0   edx: 00000001
Nov  4 15:32:13 windriver kernel: esi: cc018000   edi: c49d0b40   ebp:
cc018de0   esp: cd9f9a14
Nov  4 15:32:13 windriver kernel: ds: 0018   es: 0018   ss: 0018
Nov  4 15:32:13 windriver kernel: Process rm (pid: 2382, stackpage=cd9f9000)
Nov  4 15:32:13 windriver kernel: Stack: c19b2f1a c19b3960 00000000 000000aa
c199696e 00000000 c19afcc0 c49d
0b40
Nov  4 15:32:13 windriver kernel:        00000240 cd957188 cc018ab8 c199a2b3
c49d0b40 c192e000 c49d0b40 cc01
8000
Nov  4 15:32:13 windriver kernel:        cd957000 00000093 00000093 c49d0b40
00000018 c497be40 cd9f9adc 0000
0000
Nov  4 15:32:13 windriver kernel: Call Trace:   
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drive
rs/s+4275777306/96]
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+4275779936/96]
[st:__ins
mod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+4275661166/96]
[st:__insmod_st_O/lib/modules/2.4.21-9
9-athlon/kernel/drivers/s+4275764416/96]
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+427
5675827/96]
Nov  4 15:32:13 windriver kernel: Call Trace:    [<c19b2f1a>] [<c19b3960>]
[<c199696e>] [<c19afcc0>] [<c199a
2b3>]
Nov  4 15:32:13 windriver kernel:  
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+42756762
06/96]
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+4275676775/96]
[st:__insmod_st_O/lib/
modules/2.4.21-99-athlon/kernel/drivers/s+4275677974/96]
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kern
el/drivers/s+4275680270/96]
[st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+4275634609/96] [
st:__insmod_st_O/lib/modules/2.4.21-99-athlon/kernel/drivers/s+4275639202/96]
Nov  4 15:32:13 windriver kernel:   [<c199a42e>] [<c199a667>] [<c199ab16>]
[<c199b40e>] [<c19901b1>] [<c1991
3a2>]


Sebastian Kaps wrote:
> 
> Hi!
> 
> Is there something concerning ReiserFS I should know when migrating from
> Linux 2.4 to Linux 2.6?
> 
> I'm asking because a few days ago I downloaded and compiled 2.6.0-test9
> just for curiosity. It booted fine and I went back to my "standard"
> kernel, which was 2.4.23-pre5 at that time. Right after booting the
> 2.4.x kernel, I got lots of the following messages in my logs:
> ,----
> | kernel: ide2(33,3):vs-4080: reiserfs_free_block: free_block (2103:463529)[dev:blocknr]: bit already cleared
> `----
> The block number always differed. Gladly, "reiserfsck --rebuild-tree"
> helped to cure that, but there were lots of unusable files after that
> (e.g. parts of files overwritten with parts of other files).
> 
> Is this suppose to happen? What caused this error?
> 
> Another thing: I had to "hard reboot" my system a few times in the last
> few months. I also experienced that sometimes parts of files were
> overwritten with 0x00s or parts of other files. Can someone explain to
> me how this can happen?
> 
> --
> Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Sebastian Kaps]

Hi Christian!

> Reiserfs does only metadata journaling, if you perform a hard reset the
> data newly written to your files may be lost, because it has not been
> forced onto the disk.

The problem is not that newly written data is lost but that files
suddenly have contents that belong to other files. I could live with the
loss of unsaved data but for example I hate it when my window manager
configuration is wrecked because contents that would belong to
/var/log/messages appear in its config files. I've never had this with
ext2 but quite a few times with ReiserFS.

> If you want your data to be consistent in such cases you need to apply
> Chris Mason's data logging patches, or use the SuSE Linux kernel.
> See http://ftp.suse.com/pub/people/mason

I'll try that, thank you.

-- 
Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Vince]

Sebastian Kaps wrote:

> 
> The problem is not that newly written data is lost but that files
> suddenly have contents that belong to other files. I could live with the
> loss of unsaved data but for example I hate it when my window manager
> configuration is wrecked because contents that would belong to
> /var/log/messages appear in its config files. I've never had this with
> ext2 but quite a few times with ReiserFS.

Same here, and I couldn't agree more. I've had several bad corruption in 
my config files and logs happening because of this particular - and way 
too common - reiserfs behaviour...   :-/

>>If you want your data to be consistent in such cases you need to apply
>>Chris Mason's data logging patches, or use the SuSE Linux kernel.
>>See http://ftp.suse.com/pub/people/mason

On the same topic : does anybody know if there is a 2.6 version of these 
patches available somewhere ? (or else if a 2.6 port of those is planned 
in the short term ?)

Vincent

Re: Linux 2.4 -> 2.6 migration

[Sebastian Kaps]

Hi Jure!

> Sounds to me like you'd want to check your hardware and ram. I'm dual
> booting between redhat 2.4.20-something and 2.6.0-test9 and have no problems
> wih reiserfs partitions.

I'm pretty sure it's not hardware related. I'm running this hardware for
ages now without any problems...

-- 
Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Soeren Sonnenburg]

On Fri, 2003-11-07 at 16:09, Jure Peèar wrote:
> On Fri, 07 Nov 2003 14:39:06 +0100
> Sebastian Kaps <seb-keyword-reiserfs.0462a0@toyland.sauerland.de> wrote:
> 
> > Another thing: I had to "hard reboot" my system a few times in the last
> > few months. I also experienced that sometimes parts of files were
> > overwritten with 0x00s or parts of other files. Can someone explain to
> > me how this can happen?
> 
> Sounds to me like you'd want to check your hardware and ram. I'm dual
> booting between redhat 2.4.20-something and 2.6.0-test9 and have no problems
> wih reiserfs partitions.

Nope that is expected behaviour without data logging patches.

Soeren

Re: Linux 2.4 -> 2.6 migration

[Quinn Harris]

Did you enable IDE TCQ in your 2.6 kernel?  To quote http://www.namesys.com/
download.html
"Don't enable TCQ (Tagged Command Queuing) in 2.5 (and modified 2.4) kernels. 
TCQ is known broken currently and it will corrupt your filesystems for sure."

Quinn


On Friday 07 November 2003 06:39 am, Sebastian Kaps wrote:
> Hi!
> 
> Is there something concerning ReiserFS I should know when migrating from
> Linux 2.4 to Linux 2.6?
> 
> I'm asking because a few days ago I downloaded and compiled 2.6.0-test9
> just for curiosity. It booted fine and I went back to my "standard"
> kernel, which was 2.4.23-pre5 at that time. Right after booting the
> 2.4.x kernel, I got lots of the following messages in my logs:
> ,----
>
> | kernel: ide2(33,3):vs-4080: reiserfs_free_block: free_block
> | (2103:463529)[dev:blocknr]: bit already cleared
>
> `----
> The block number always differed. Gladly, "reiserfsck --rebuild-tree"
> helped to cure that, but there were lots of unusable files after that
> (e.g. parts of files overwritten with parts of other files).
> 
> Is this suppose to happen? What caused this error?
> 
> Another thing: I had to "hard reboot" my system a few times in the last
> few months. I also experienced that sometimes parts of files were
> overwritten with 0x00s or parts of other files. Can someone explain to
> me how this can happen?
> 
> -- 
> Ciao, Sebastian
> 

Re: Linux 2.4 -> 2.6 migration

[Alex Zarochentsev]

On Fri, Nov 07, 2003 at 02:39:06PM +0100, Sebastian Kaps wrote:
> Hi!
> 
> Is there something concerning ReiserFS I should know when migrating from
> Linux 2.4 to Linux 2.6?
> 
> I'm asking because a few days ago I downloaded and compiled 2.6.0-test9
> just for curiosity. It booted fine and I went back to my "standard"
> kernel, which was 2.4.23-pre5 at that time. Right after booting the
> 2.4.x kernel, I got lots of the following messages in my logs:
> ,----
> | kernel: ide2(33,3):vs-4080: reiserfs_free_block: free_block (2103:463529)[dev:blocknr]: bit already cleared

test-9/reiser3 works fine for us.  Your problems could be due to a bug in
updated IDE driver, for example.  Do other filesystems (like ext2) work? 

Can you send a .config (from 2.6.0-test-9) and describe your hardware configuration
to the reiserfs-dev@namesys.com mailing list?

> `----
> The block number always differed. Gladly, "reiserfsck --rebuild-tree"
> helped to cure that, but there were lots of unusable files after that
> (e.g. parts of files overwritten with parts of other files).

I think that reiserfsck --rebuild-tree trusts the bitmap content, which is not
good for your case.  Vitaly will answer better.

> 
> Is this suppose to happen? What caused this error?
> 
> Another thing: I had to "hard reboot" my system a few times in the last
> few months. I also experienced that sometimes parts of files were
> overwritten with 0x00s or parts of other files. Can someone explain to
> me how this can happen?
> 
> -- 
> Ciao, Sebastian

-- 
Alex.

Re: Linux 2.4 -> 2.6 migration

[Vitaly Fertman]

On Saturday 08 November 2003 13:11, Alex Zarochentsev wrote:
> On Fri, Nov 07, 2003 at 02:39:06PM +0100, Sebastian Kaps wrote:
> > Hi!
> >
> > Is there something concerning ReiserFS I should know when migrating from
> > Linux 2.4 to Linux 2.6?
> >
> > I'm asking because a few days ago I downloaded and compiled 2.6.0-test9
> > just for curiosity. It booted fine and I went back to my "standard"
> > kernel, which was 2.4.23-pre5 at that time. Right after booting the
> > 2.4.x kernel, I got lots of the following messages in my logs:
> > ,----
> >
> > | kernel: ide2(33,3):vs-4080: reiserfs_free_block: free_block
> > | (2103:463529)[dev:blocknr]: bit already cleared
>
> test-9/reiser3 works fine for us.  Your problems could be due to a bug in
> updated IDE driver, for example.  Do other filesystems (like ext2) work?
>
> Can you send a .config (from 2.6.0-test-9) and describe your hardware
> configuration to the reiserfs-dev@namesys.com mailing list?
>
> > `----
> > The block number always differed. Gladly, "reiserfsck --rebuild-tree"
> > helped to cure that, but there were lots of unusable files after that
> > (e.g. parts of files overwritten with parts of other files).
>
> I think that reiserfsck --rebuild-tree trusts the bitmap content, which is
> not good for your case.  Vitaly will answer better.

Bitmaps do not say that this block belongs to some particular file,
so bitmaps have nothing to do with the described problem when 
parts of files overwritten with parts of other files. But when there
are some blocks in use that have been already freed there could 
be another problem -- parts of files are lost, there are just zeroes 
in these parts, or some files are lost after fsck. Probably scanning
through all blocks of the partition with --scan-whole-partition 
option would help here, but this option should be used with 
extreme caution as you can run out of the partition space with it.

-- 
Thanks,
Vitaly Fertman

Re: Linux 2.4 -> 2.6 migration

[Sebastian Kaps]

Hi Alex!

> test-9/reiser3 works fine for us.  Your problems could be due to a bug in
> updated IDE driver, for example.  Do other filesystems (like ext2) work? 

I only have one big reiserfs partition, so I can't test it.

> Can you send a .config (from 2.6.0-test-9) and describe your hardware
> configuration to the reiserfs-dev@namesys.com mailing list?

Sorry, I did an "rm -rf" on the 2.6.x tree two days ago. I guess that
enabling TCQ was the mistake, as Quinn Harris pointed out.

-- 
Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Matthias Andree]

Sebastian Kaps <seb-keyword-reiserfs.0462a0@toyland.sauerland.de> writes:

> Is there something concerning ReiserFS I should know when migrating from
> Linux 2.4 to Linux 2.6?

No. Make sure the file system is fine before booting a new kernel, using
a CURRENT reiserfsprogs is essential. Namesys have made lots of fixes
recently.

-- 
Matthias Andree

Encrypt your mail: my GnuPG key ID is 0x052E7D95

Re: Linux 2.4 -> 2.6 migration

[Sebastian Kaps]

Hi Quinn!

> Did you enable IDE TCQ in your 2.6 kernel?  To quote http://www.namesys.com/
> download.html

I guess I did enable TCQ.

> "Don't enable TCQ (Tagged Command Queuing) in 2.5 (and modified 2.4) kernels. 
> TCQ is known broken currently and it will corrupt your filesystems for sure."

Okay, that would explain it...

-- 
Ciao, Sebastian

Re: Linux 2.4 -> 2.6 migration

[Alejandro Sanchez Acosta]

El s?, 08-11-2003 a las 05:08, Quinn Harris escribió:
> Did you enable IDE TCQ in your 2.6 kernel?  To quote http://www.namesys.com/
> download.html
> "Don't enable TCQ (Tagged Command Queuing) in 2.5 (and modified 2.4) kernels. 
> TCQ is known broken currently and it will corrupt your filesystems for sure."

What does it do Tagged Command Queuing?

Thanks in advance.
-- 
Alejandro Sanchez Acosta <alejandro@hispafuentes.com>

Re: Linux 2.4 -> 2.6 migration

[Marcelo Pacheco]

On Sunday 09 November 2003 03:45, Alejandro Sanchez Acosta wrote:
> El s?, 08-11-2003 a las 05:08, Quinn Harris escribió:
> > Did you enable IDE TCQ in your 2.6 kernel?  To quote
> > http://www.namesys.com/ download.html
> > "Don't enable TCQ (Tagged Command Queuing) in 2.5 (and modified 2.4)
> > kernels. TCQ is known broken currently and it will corrupt your
> > filesystems for sure."
>
> What does it do Tagged Command Queuing?
>
> Thanks in advance.
TCQ is a feature that allows a host to send multiple commands to IDE devices 
and have them queued and executed in the most efficient manner to the device, 
being interrupted as they become ready. This is a standard SCSI feature that 
is available on higher featured drives. Lots of drives don't have this 
feature, but once it's stable on Linux, it should be ok to enable it by 
default and let Linux disable it for devices that don't support it. However 
today it's not ready for production.

Marcelo Pacheco