Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

[jbingemer]

[-- Attachment #1: Type: text/plain, Size: 842 bytes --]

Yury Umanets <umka@namesys.com> schrieb am 21.07.2003 12:58:23:


> > Jul  6 02:01:47 bi kernel: is_tree_node: node level 17747 does not
> > match to the expected one 1
> > Jul  6 02:01:47 bi kernel: vs-5150: search_by_key: invalid format
> > found in block 42752. Fsck?
> Yes, this definitely means, that you have corrupted some formatted nodes
> on your reiserfs partition.

This does not sound very good.


> > Just to be sure nothing is missing, I uploaded the output of dmesg to
> > http://www.baetmaen.de/dmsg.tar.gz (6kb)
> Sorry, but we need system log. Something like /var/log/messages. 

Oh, I wanted to give you some information on the hardware in case I missed 
something.

OK here is the syslog, I have no idea on how to sort out relevant 
information, sorry about that.

http://www.baetmaen.de/syslog.tar.gz (40K)

Regards
joerg

Re: Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

[Yury Umanets]

On Mon, 2003-07-21 at 15:10, jbingemer@oilgear.de wrote:
> Yury Umanets <umka@namesys.com> schrieb am 21.07.2003 12:58:23:
> 
> 
> > > Jul  6 02:01:47 bi kernel: is_tree_node: node level 17747 does not
> > > match to the expected one 1
> > > Jul  6 02:01:47 bi kernel: vs-5150: search_by_key: invalid format
> > > found in block 42752. Fsck?
> > Yes, this definitely means, that you have corrupted some formatted nodes
> > on your reiserfs partition.
> 
> This does not sound very good.
> 
> 
> > > Just to be sure nothing is missing, I uploaded the output of dmesg to
> > > http://www.baetmaen.de/dmsg.tar.gz (6kb)
> > Sorry, but we need system log. Something like /var/log/messages. 
> 
> Oh, I wanted to give you some information on the hardware in case I missed 
> something.
> 
> OK here is the syslog, I have no idea on how to sort out relevant 
> information, sorry about that.
> 
> http://www.baetmaen.de/syslog.tar.gz (40K)

Okay, syslog does not contain IO errors, but contains lots of messages
about formatted nodes corrupted. You definitely need to run fsck. 

What if you try to read that blocks fsck has hanged on them by dd for
instance?

For instance:
dd if=/dev/hda1 of=/tmp/block bs=4096 count=1 skip=BLOCK_NUMBER
> 
> Regards
> joerg
-- 
We're flying high, we're watching the world passes by...

Re: Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

[Yury Umanets]

On Mon, 2003-07-21 at 15:10, jbingemer@oilgear.de wrote:
> Yury Umanets <umka@namesys.com> schrieb am 21.07.2003 12:58:23:
> 
> 
> > > Jul  6 02:01:47 bi kernel: is_tree_node: node level 17747 does not
> > > match to the expected one 1
> > > Jul  6 02:01:47 bi kernel: vs-5150: search_by_key: invalid format
> > > found in block 42752. Fsck?
> > Yes, this definitely means, that you have corrupted some formatted nodes
> > on your reiserfs partition.
> 
> This does not sound very good.
> 
> 
> > > Just to be sure nothing is missing, I uploaded the output of dmesg to
> > > http://www.baetmaen.de/dmsg.tar.gz (6kb)
> > Sorry, but we need system log. Something like /var/log/messages. 
> 
> Oh, I wanted to give you some information on the hardware in case I missed 
> something.
> 
> OK here is the syslog, I have no idea on how to sort out relevant 
> information, sorry about that.
> 
> http://www.baetmaen.de/syslog.tar.gz (40K)

Yet another issue. Can you run badblocks command in maner like the
following:

badblocks -b 4096 -o /tmp/badblocks /dev/hde

And send us /tmp/badblocks.
> 
> Regards
> joerg
-- 
We're flying high, we're watching the world passes by...

Antwort: Re: Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

[jbingemer]

[-- Attachment #1: Type: text/plain, Size: 1072 bytes --]

Yury Umanets <umka@namesys.com> schrieb am 21.07.2003 13:33:19:


I just took a look at the syslog and found the following line:

Jul  9 18:36:20 bi kernel:  I/O error: dev 0b:01, sector 0
Jul 21 01:20:13 bi kernel: hde: dma_intr: error=0x40 { UncorrectableError 
}, LBAsect=4903084, sector=4902920
Jul 21 01:20:13 bi kernel: end_request: I/O error, dev 21:05 (hde), sector 
4902920
bi:~ #

> Yet another issue. Can you run badblocks command in maner like the
> following:
> 
> badblocks -b 4096 -o /tmp/badblocks /dev/hde
> 
> And send us /tmp/badblocks.

I think the time it takes to run this command depends on how many bad 
blocks are found, but how can I check if it is still active/doing 
something ?

When I started the command, top reported 33% CPU time used by badblocks, 
but after 2 minutes or so it stops taking cputime.

  918 root      15   0  1668 1668  1260 S     0.0  0.6   0:00 bash
  937 root      15   0   472  472   340 D     0.0  0.1   0:20 badblocks
 1044 root      15   0  2088 2088  1824 S     0.0  0.8   0:00 sshd

Is this normal ?

Regards
joerg

Re: Antwort: Re: Antwort: Re: Antwort: Re: Problems with reiserfsck 3.6.9 (Newbie)

[Yury Umanets]

On Mon, 2003-07-21 at 16:35, jbingemer@oilgear.de wrote:
> Yury Umanets <umka@namesys.com> schrieb am 21.07.2003 13:33:19:
> 
> 
> I just took a look at the syslog and found the following line:
> 
> Jul  9 18:36:20 bi kernel:  I/O error: dev 0b:01, sector 0
> Jul 21 01:20:13 bi kernel: hde: dma_intr: error=0x40 { UncorrectableError 
> }, LBAsect=4903084, sector=4902920
> Jul 21 01:20:13 bi kernel: end_request: I/O error, dev 21:05 (hde), sector 
> 4902920
> bi:~ #
This probably explains problems with hanging fsck. It was unable to read
some blocks from your /dev/hde.
> 
> > Yet another issue. Can you run badblocks command in maner like the
> > following:
> > 
> > badblocks -b 4096 -o /tmp/badblocks /dev/hde
> > 
> > And send us /tmp/badblocks.
> 
> I think the time it takes to run this command depends on how many bad 
> blocks are found, but how can I check if it is still active/doing 
> something ?
Actually we just need to know, is there some bad blocks or not. So, it
is not neccessary to wait until finish.
> 
> When I started the command, top reported 33% CPU time used by badblocks, 
> but after 2 minutes or so it stops taking cputime.
> 
>   918 root      15   0  1668 1668  1260 S     0.0  0.6   0:00 bash
>   937 root      15   0   472  472   340 D     0.0  0.1   0:20 badblocks
>  1044 root      15   0  2088 2088  1824 S     0.0  0.8   0:00 sshd
> 
> Is this normal ?
Yes, why not. It is definitely IO-bounds process, so, the most of its
time will be spend for waiting on IO operation.


Okay, so, it seems, that your drive has badblocks. If so, I'll give you
instructions, which will help to make backup and repair your reiserfs
partition. Can we do so according to http://www.namesys.com/support.html
terms?

Thanks.
> 
> Regards
> joerg
-- 
We're flying high, we're watching the world passes by...