UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY and Harddisk problem

[David Tonhofer, m-plify S.A.]

--On Wednesday, March 22, 2006 8:21 AM -0500 "Bliss, Aaron" <ABliss at preferredcare.org> wrote:

> What can I do for such a case? Can't Linux mark bad blocks so that they
> can't be used? Does it have something like scandisk or chkdsk?

There is 'badblocks' which can destructively or non-destructively
look for bad blocks. If your harddisk is seriously damaged though
it will find a LOT of these and it may take literally days until it
gets to the end of the check (I tried that once - found that the first
N and last M blocks had a high error probability but block 0 was
still good; then set up the disk so that these blocks were not in
any used partition. Last week the disk died soundlessly. Only do
this stunt if you have very good backup and lots of time to spend)

Modern disks should register & remap (a few) bad blocks by themselves
I think.

You can feed the output of badblocks back into mke2fs so that
these blocks are avoided during fs construction.	

> Don't tell me that the harddisk cannot be used anymore cause I have
> another harddisk which is less than 2 years old since I bought and it
> has some such problems. Luckily, it can still be used. In a day, I use
> less than 8 hours and I don't use the computer everyday. (Average of 3
> hrs a day)

If it's under guarantee you can probably exchange it under the
manufacturer's exchange program. If it's a Fujitsu, check that it
isn't one of the 'bad series' that has known defects. Disks can run
24/7 without trouble. Power cycles are more damaging to them than
continuous running IMHO.

Good luck

 -- David