disk inconsistency - should 'fsck' always executed with -y ?

[ddm at pizzashack.org (Derek Martin)]

Today, Glenn Burkhardt gleaned this insight:

> > If that file system is messed up, running fsck with the -y option could 
> > cause a lot of data to be lost. Even when the file system is corrupted, 
> > you might be able to copy off the data. 
> 
> Do you know of any good references on how to do that, to look for potentially
> lost data?

Usually when files are deleted, they are put in the lost+found directory
in the top level directory of the filesystem (NEVER delete that directory
on an HP-UX machine!  You will not be able to run fsck, as I found out the
hard way!  IIRC Linux (or I guess I should say ext2) just re-creates the
directory if it's missing.)

Just because the file is there though, doesn't mean you'll be able to
figure out what it was...  There are also probably cases where the
filesystem was so out of whack that some data will just be lost.

I haven't come across any references on how to really pick apart a bad
filesystem.. you'd probably need to be intimately familiar with the
filesystem implementation, and pick it apart manually, block by
block.  Not something I ever want to do.  Restoring from back-ups is much
much easier.

> > It is simply when there is a serious problem, you might want 
> > a human to take a look at it. 
> 
> Again, just how does one do that?

For the Average Mortal(tm), this usually just amounts to watching fsck
run, which may or may not give you some idea of how bad the damage was,
and possibly what might have been lost.  If nothing else, it might give
you an idea whether or not you should restore from your latest back-up
(assuming you do back-ups)....

-- 
Derek Martin
ddm at pizzashack.org

-
Subcription/unsubscription/info requests: send e-mail with
"subscribe", "unsubscribe", or "info" on the first line of the
message body to discuss-request at blu.org (Subject line is ignored).