I'm using XFS for a very long time under Linux (was an Irix sysadmin
too), but now I have some serious questions.
One of my disks is failing, a 9Gb IBM SCSI disk. It seems that only
the first some sectors have read-errors, including the primary
superblock. :-( So I have imaged the disk with dd to a file, and run
xfs_repair on it. It said:
Phase 1 - find and verify superblock...
bad primary superblock - bad magic number !!!
attempting to find secondary superblock...
It found eight secondary superblocks, with the following messages:
found candidate secondary superblock...
unable to verify superblock, continuing...
What does verify superblock means? I think it checks some offsets, that
may not be correct because dd does not copy sectors with errors; so some
sectors are missing from the beginning, moving all offsets closer to the
beginning of the image. Anyone can confirm this? So all I may do is to
find out this offset and put the exact number of bytes at the beginning
of the image, right?
Information about the system:
- it's a PIII with onboard SCSI controller,
- disk is an IBM SCSI one, 9 Gb, with only one partition with XFS,
formatted under 2.6 AFAICR,
- current kernel is 2.6.12, under Debian Sarge, so xfsprogs is 2.6.20.
What other things should I try out?
Thanks for any pointers in advance,
BorsodChem Joint-Stock Company www.debian.org Linux Support Center
Software engineer Debian Developer Developer
Description: This is a digitally signed message part