Re: a subtle bug of ext2 or e2fsck ? (fwd)

Theodore Y. Ts'o (tytso@mit.edu)
Sun, 19 Sep 1999 07:30:32 -0400


Date: Sat, 18 Sep 1999 17:14:59 +1000 (EST)
From: Hao He <he@Physics.usyd.edu.au>

I recently have had some troubles with my file systems. e2fsck (1.14 )
reports ``Error reading block #### (Attempt to read block from filesystem
resulted in short read) while doing inode scan. Ignore error<y>". I
answered 'y'. However, I got more and more such errors. Eventually, I got
''Attempt to read black from filesystem resulted in short read while
trying to open /dev/hda3''.

I have three partitions from dev/hda1 to /dev/hda3 and I have the same
problems with all three partitions. I reformatted dev/hda1 and it seems
fine. So this is not a hardware problem.

The short read error means that the kernel returned an error when e2fsck
tried to read a block. This usually means a hardware error, but it can
mean that the partition table got changed so that the partition is
simaller than the logical partition size. You can check to see if there
were I/O errors by looking in /var/log/messages for any errors reported
by the IDE disk driver.

As for why /dev/hda1 worked after you reformatted it, there are two
possibilities. The first is that the partition table got corrupted, and
so when you reformatted the filesystem, you made the new filesystem
which was properly sized for the new partition size, and so the problems
went away. The second is that when you rewrote the block, the IDE disk
automatically rerouted the filed block to a new block from its pool of
spare blocks.

I was too stupid to backup my files.

I'll refrain from making the obvious comment.... :-)

- Ted

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/