Messages in this thread | | | Date | Wed, 9 Jul 1997 13:53:17 -0400 | From | "Theodore Y. Ts'o" <> | Subject | Re: corrupt block makes file system repair impossible? |
| |
Date: Wed, 9 Jul 1997 17:53:22 +0200 (MET DST) From: Thomas Koenig <ig25@mvmap66.ciw.uni-karlsruhe.de>
I just tried to repair a file system on a disk with some bad sectors, and e2fsck told me:
Block 4 in the primary group descriptors is on the bad block list
If the block is really bad, the filesystem can not be fixed. You can clear the this block from the bad block list and hope that block is really OK, but there are no guarantees.
[...]
Is this true? What's the point in having multiple superblocks if trashing a single sector can still completely wreck your file system?
It's at least partially true. The kernel currently assumes that the block group descriptors follow the superblock. If you specify another superblock, it'll use another set of block group descriptors, though.
However, using an alternate superblock for a filesystem has typically been viewed as a stopgap measure, instead of a something you would do as a long-term thing. For example, there's no way to make fsck call e2fsck with the right arguments to use an alternate superblock as part of the default boot sequence for a particular filesystem (based on some fields in fstab, for example). You can specify the alternate superblock if you run e2fsck manually, and it's currently primarily used to recover from problems where your primary superblock has been overwritten. But the tools aren't currently set up to make this very easy.
- Ted
| |