I am just in deep trouble with a new Linux installation on a 120 MB IDE
drive. This drive has been used for some years now and has acquired some
defective sectors in the meantime. (Of course I know it should detect its
errors and correct them itself - I just suppose there's no space left for
further repairing procedures).
If I issue
mke2fs -c ...
the filesystem is created and *some* bad blocks are detected. My problem is
that there seem to be some "light" bad blocks that aren't found by the
(probably quite sloppy?) checking procedure of mkfs. Linux dies quietly
after installation with "bad Inode" errors.
The same problem occurs with the swap partition.
So what I need is one of the following:
- a HDD checking tool that creates a (thoroughly tested) bad block list
and a description of the mkfs -l <bad_block_list> format
- the mke2fs/mkswap sources (I will just be able to tighten the testing
loops and make the whole thing more thorough...)
- or some other ingenious tip (short of "trash this drive"...)