* How to recover data with a dead journal?
@ 2002-09-06 12:13 Matthew Bloch
2002-09-06 12:17 ` Oleg Drokin
0 siblings, 1 reply; 6+ messages in thread
From: Matthew Bloch @ 2002-09-06 12:13 UTC (permalink / raw)
To: reiserfs-list
Here's a problem; my drive is on its way out but the latest corrupt sector is
the ReiserFS journal (from the ReiserFS in kernel 2.4.18):
factotum:/home/mattbee# mount /dev/hdb3 /old/
reiserfs: checking transaction log (device 03:43) ...
hdb: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hdb: read_intr: error=0x40 { UncorrectableError }, LBAsect=733664,
sector=40160
end_request: I/O error, dev 03:43 (hdb), sector 40160
hdb: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hdb: read_intr: error=0x40 { UncorrectableError }, LBAsect=750560,
sector=57056
end_request: I/O error, dev 03:43 (hdb), sector 57056
journal-1212: REPLAY FAILURE fsck required! buffer write failed
Replay Failure, unable to mount
reiserfs_read_super: unable to initialize journal space
mount: wrong fs type, bad option, bad superblock on /dev/hdb3,
or too many mounted file systems
Of course I've fscked it while telling it to ignore the journal, and it found
a few errors but it still doesn't change the fact that mount won't mount
without a journal. mount's nolog option doesn't seem to be implemented yet--
is there a later version that does implement it, and is this the only way to
get at what's left of my data?
I've tried:
dd if=/dev/hdb3 of=deaddrive conv=noerror
mount deaddrive /old/ -t reiserfs -o loop
but mount reports, somewhat cryptically that /old isn't a directory, which it
plainly is :)
I only want to read the remnants off once then discard the drive; any
suggestions for the quicket solution?
--
Matthew Bloch Bytemark Computer Consulting Limited
http://www.bytemark.co.uk/
tel. +44 (0) 8707 455026
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: How to recover data with a dead journal?
2002-09-06 12:13 How to recover data with a dead journal? Matthew Bloch
@ 2002-09-06 12:17 ` Oleg Drokin
2002-09-06 12:43 ` Matthew Bloch
0 siblings, 1 reply; 6+ messages in thread
From: Oleg Drokin @ 2002-09-06 12:17 UTC (permalink / raw)
To: Matthew Bloch; +Cc: reiserfs-list
Hello!
On Fri, Sep 06, 2002 at 01:13:24PM +0100, Matthew Bloch wrote:
> Here's a problem; my drive is on its way out but the latest corrupt sector is
> the ReiserFS journal (from the ReiserFS in kernel 2.4.18):
Usually we deal with HW problems based on http://namesys.com/support.html
terms.
Similar questions were answered on this list already, though, so you might
want to check the archive.
> dd if=/dev/hdb3 of=deaddrive conv=noerror
dd still corrupts data in such a case.
> mount deaddrive /old/ -t reiserfs -o loop
> but mount reports, somewhat cryptically that /old isn't a directory, which it
> plainly is :)
Data is shifted, so FS structure is destroyed.
Bye,
Oleg
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: How to recover data with a dead journal?
2002-09-06 12:17 ` Oleg Drokin
@ 2002-09-06 12:43 ` Matthew Bloch
2002-09-06 13:43 ` Dieter Nützel
0 siblings, 1 reply; 6+ messages in thread
From: Matthew Bloch @ 2002-09-06 12:43 UTC (permalink / raw)
To: reiserfs-list
On Friday 06 September 2002 13:17, Oleg Drokin wrote:
> Hello!
>
> On Fri, Sep 06, 2002 at 01:13:24PM +0100, Matthew Bloch wrote:
> > Here's a problem; my drive is on its way out but the latest corrupt
> > sector is the ReiserFS journal (from the ReiserFS in kernel 2.4.18):
>
> Usually we deal with HW problems based on http://namesys.com/support.html
> terms.
> Similar questions were answered on this list already, though, so you might
> want to check the archive.
Fair enough; I thought there might be some experienced users here who could
help as well. I'm trying dd with a different set of options to try to make
it leave gaps where it can't read, so hopefully that should sort it. If it
doesn't I'll put a few dollars your way and make an official support request
out of it.
cheers,
--
Matthew Bloch Bytemark Computer Consulting Limited
http://www.bytemark.co.uk/
tel. +44 (0) 8707 455026
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: How to recover data with a dead journal?
2002-09-06 12:43 ` Matthew Bloch
@ 2002-09-06 13:43 ` Dieter Nützel
2002-09-06 13:55 ` Matthew Bloch
0 siblings, 1 reply; 6+ messages in thread
From: Dieter Nützel @ 2002-09-06 13:43 UTC (permalink / raw)
To: Matthew Bloch, reiserfs-list
On Friday 06 September 2002 14:43, Matthew Bloch wrote:
> On Friday 06 September 2002 13:17, Oleg Drokin wrote:
> > Hello!
> >
> > On Fri, Sep 06, 2002 at 01:13:24PM +0100, Matthew Bloch wrote:
> > > Here's a problem; my drive is on its way out but the latest corrupt
> > > sector is the ReiserFS journal (from the ReiserFS in kernel 2.4.18):
> >
> > Usually we deal with HW problems based on http://namesys.com/support.html
> > terms.
> > Similar questions were answered on this list already, though, so you
> > might want to check the archive.
>
> Fair enough; I thought there might be some experienced users here who could
> help as well. I'm trying dd with a different set of options
NO, NO, NO!
Read what Oleg told you and READ the archive!
First:
Start with only one disk (IDE/ATA) per channel!
Do not enable "auto detection" in the BIOS.
Let the Linux IDE kernel/module driver do the work.
Try not to mount the broken disk.
Only have a look at the partition tables (fdisk/cfdisk).
Use "dd_rescue" (created by SuSE) and copy the whole disk to a new empty disk
of the same size or maybe bigger (cold take ages, 20~25 hours for a 20/40/60
GB disk, bigger disks or more errors even longer). Or make "file" copies to a
bigger disk so that you can mount with the loop device.
Make a second copy of the "good" disk (takes 1-2 hours at raw disk speed) and
then work on it. If you damage some things you have your second "good" one.
> to try to make
> it leave gaps where it can't read, so hopefully that should sort it.
dd_rescue do this automatically.
> If it doesn't I'll put a few dollars your way and make an official support
> request out of it.
>
> cheers,
Better after your work...;-)
Regards,
Dieter
--
Dieter Nützel
Graduate Student, Computer Science
University of Hamburg
Department of Computer Science
@home: Dieter.Nuetzel at hamburg.de (replace at with @)
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: How to recover data with a dead journal?
2002-09-06 13:43 ` Dieter Nützel
@ 2002-09-06 13:55 ` Matthew Bloch
2002-09-06 16:00 ` Valdis.Kletnieks
0 siblings, 1 reply; 6+ messages in thread
From: Matthew Bloch @ 2002-09-06 13:55 UTC (permalink / raw)
To: reiserfs-list
On Friday 06 September 2002 14:43, Dieter Nützel wrote:
> On Friday 06 September 2002 14:43, Matthew Bloch wrote:
> > On Friday 06 September 2002 13:17, Oleg Drokin wrote:
> > > Hello!
> > >
> > > On Fri, Sep 06, 2002 at 01:13:24PM +0100, Matthew Bloch wrote:
> > > > Here's a problem; my drive is on its way out but the latest corrupt
> > > > sector is the ReiserFS journal (from the ReiserFS in kernel 2.4.18):
> > >
> > > Usually we deal with HW problems based on
> > > http://namesys.com/support.html terms.
> > > Similar questions were answered on this list already, though, so you
> > > might want to check the archive.
> >
> > Fair enough; I thought there might be some experienced users here who
> > could help as well. I'm trying dd with a different set of options
>
> NO, NO, NO!
> Read what Oleg told you and READ the archive!
I read what he said about dd, so I tried dd with bs=512, sync and noerror
options, and the image it created mounted perfectly (because 'sync' ensured
that holes are left, and bs=512 meant that the holes were as small as they
could be). It's only a small disc so it didn't take long. I'll remember
dd_rescue for next time, though disc recovery in general is dull and
potentially heartbreaking ;-) so I think I'll just splash out on a mirror
drive so there is no next time!
thanks,
--
Matthew Bloch Bytemark Computer Consulting Limited
http://www.bytemark.co.uk/
tel. +44 (0) 8707 455026
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: How to recover data with a dead journal?
2002-09-06 13:55 ` Matthew Bloch
@ 2002-09-06 16:00 ` Valdis.Kletnieks
0 siblings, 0 replies; 6+ messages in thread
From: Valdis.Kletnieks @ 2002-09-06 16:00 UTC (permalink / raw)
To: Matthew Bloch; +Cc: reiserfs-list
[-- Attachment #1: Type: text/plain, Size: 1056 bytes --]
On Fri, 06 Sep 2002 14:55:46 BST, Matthew Bloch <matthew@bytemark.co.uk> said:
> could be). It's only a small disc so it didn't take long. I'll remember
> dd_rescue for next time, though disc recovery in general is dull and
> potentially heartbreaking ;-) so I think I'll just splash out on a mirror
> drive so there is no next time!
I've had to help recover a large RAID5 with an Oracle database on it after
the controller went bonkers and flushed its cache to the correct block
numbers on the incorrect disk numbers. Ouch. At least it wasn't like
the one shop I know of that bought a terabyte of RAID-5 and no backup
device because it was hot-swap RAID-5 with hot spares - worked great till
a plumbing leak on the floor above dumped lots of water into the cabinet...
Take regular backups anyhow - the Universe runs by the laws of thermodynamics,
and is continually trying to increase entropy. Your data has a big bullseye
target on it. ;)
--
Valdis Kletnieks
Computer Systems Senior Engineer
Virginia Tech
[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2002-09-06 16:00 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-09-06 12:13 How to recover data with a dead journal? Matthew Bloch
2002-09-06 12:17 ` Oleg Drokin
2002-09-06 12:43 ` Matthew Bloch
2002-09-06 13:43 ` Dieter Nützel
2002-09-06 13:55 ` Matthew Bloch
2002-09-06 16:00 ` Valdis.Kletnieks
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.