I think I must have had a head crash or something (WDC Caviar 21000,
never again). The kernel started printing out { unrecoverable Error } or
somesuch, and indeed, e2fsck -c found the offending blocks. Subsequent
runs did not though - have these blocks been remapped? Is my data good?
e2fsck still reports that there are errors (haven't dared run it in r/w
mode yet!).
How good is the badblocks program? I don't know whether it was
pretty-user-interface-psychology or not, but I always felt better doing
a scan with Norton's CALIBRAT or Disk Doctor than when I run e2fsck
-c.... it used to run (what seemed like) detailed tests of the disk
controller etc.. Can we do this on Linux, or was it just B.S.?
How about some program that liaises with the IDE driver in doing heavy
diagnostics - in fact, why not have a full diagnostic suite for Linux?
I have a feeling that as Windows progresses through its various
revisions, the efficacy of such diagnostic software decreases. Older
versions of diag software could get in at a lower level, methinks. Plus,
you have to have more and more heavy OS running before you can run your
diags. Run Norton Disk Doctor on a possibly-damaged filesystem? Sure,
but you have to have Windows running first, and heaven knows what
that'll do to your FS before you can get it to a state where you can run
NDD! At least when it was DOS, we knew what it was doing...
I suppose there was just less to go wrong, back then. A recent review of
diagnostic tools in a UK PC magazine said that "most users wouldn't know
what to do with a hex dump of their FAT" (or words to that effect) - so
you can expect Norton Disk Editor to vanish from the next version.
Surely, this is dumbing down: it might be acceptable in the Windows
world, but we're not like that.
Is this a relevant discussion on linux-kernel? Not sure, but it needs
low-level access, and isn't the kernel best placed for this? Sorry if
I'm just wasting your bandwidth.
Looking forward to running "The Torvalds Utilities" v1.00!
Cheers,
Matt
-- Matt J. Gumbley- 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/