Hello,

I have experienced this very same problem and saw several posts about it,
but no responses. I thought I would share my method for getting out of
this mess.

Since I was totally unable to mount, or fsck this reiserfs parition, I
thought I was totally hosed, except for the fact that running
debugreiserfs told me that there was data on there. What I did was put a
fresh, unformatted disk on the system that was larger than my hosed drive.
The new disk just had a partition put on it, no file system. Then I just
ran dd from and to the raw devices, in this case, my RAID partition was
hosed, /dev/rd/c0d0p1 (Mylex block device) and I had a new 18-gig scsi
drive, /dev/sdb1. After running "dd=if=/dev/rd/c0d0p1 of=/dev/sdb1" which
took about an hour and half, I was able to cleanly mount /dev/sdb1, and
all the data was totally pristine -- a reiserfsck returned no errors.

I think there is a bug somewhere that allows this to happen, odd thing is,
the drive was only 34% full. If anyone wants to look at some debug dumps,
I've got them.

I hope this helps some people get their data back (and keep their jobs!)

Andreas Stollar
SPEAKEASY.net System Administrator


--snip--
Here's the exact message:
-----------
root@deot:~/reiserfsprogs-3.x.0j/fsck > ./reiserfsck /dev/hdb6

<-------------reiserfsck, 2001------------->
reiserfsprogs 3.x.0j

Will read-only check consistency of the partition
Will put log info to 'stderr'
Do you want to run this program?[N/Yes] (note need to type Yes):Yes
Analyzing journal..last flushed trans 277332, mount_id 48, will replay
from 277333 up to 277355:Yes?Mount_id 48, transaction 277333, desc block
5848, commit block 5911: ( 311179 294912 16 311178 25312 188757 86595
311280bwrite: write 4096 bytes returned -1 (block=311280, dev=3): No space
left on device
----------

I saw that there's a switch --no-journal-replays, but it doesn't seem to
work on my version, which seems to be the latest. I'll really appreciate
your help.

Thanks in advance!!

Cya,
Oren.

Reply via email to