i found out that the last dd i did was incomplete, so i did a new one
(which took over 24h) and there were more bad blocks than i thought
before. about 1-2% of the hdd are bad blocks.
i got xfs_check to work now (xfs_ncheck still segfaults though), but things
don't look
too got. it complains on every block:
...
block 10/5704 out of range
block 10/5705 out of range
block 10/5706 out of range
...
what does that mean? it's not mentioned in the manpage. xfs_repair still
doesn't work, finds many secondary superblocks but can't verify them.
when i try to mount the fs i get:
Jul 23 15:17:40 knecht kernel: XFS mounting filesystem ide0(3,5)
Jul 23 15:17:40 knecht kernel: XFS: Log inconsistent or not a log (last==0,
first!=1)
Jul 23 15:17:40 knecht kernel: XFS: empty log check failed
Jul 23 15:17:40 knecht kernel: XFS: log mount/recovery failed
Jul 23 15:17:40 knecht kernel: XFS: log mount failed
anyone got some ideas what i can do? i thought maybe the dd options
weren't correct, i used: dd conv=noerror </dev/hdc2 >/dev/hda5
now i found the following in the dd manpage:
sync pad every input block with NULs to ibs-size; when
used
maybe that'd be a good idea. what is the standard xfs blocksize, 4096
no? that'd be dd conv=noerror,sync bs=4096 </dev/hdc2 >/dev/hda5
it would actually help me a lot already if i'd just get the filenames
and directory structure out of the partition, is there some way to do
this besides xfs_ncheck?
can anyone comment on this? help much appreciated :)
thomas
|