Subject: Re: problem with wd error handling?
To: Charles M. Hannum <abuse@spamalicious.com>
From: Matthias Drochner <M.Drochner@fz-juelich.de>
List: tech-kern
Date: 08/15/2004 17:26:17
abuse@spamalicious.com said:
> Many modern drives still use the AMNF bit.
Thanks -- so this looks serious, and I should get a new disk...
> > Can I assume that
> > the block is already remapped internally by the drive?
> You can hope.
I'l try to write to that sector; if it succeeds i know it is.
Have to find out which file is affected before so I know what
gets corrupted. What is the easiest way for this?
(find -exec cat {} and watching the error messages should work
hopefully, as long as it isn't a directory)
[dkctl]
> It doesn't show you the drive's remapping table.
I see - it is just IDE... And it probably doesn't make sense for
the driver to add an entry on a read error - just on a non-
recoverable write error.
ragge@ludd.luth.se said:
> Doesn't bad144 work on wd-type drives anymore?
It seems to pretend that it works in principle... I'm afraid
the sectors were not reserved on filesystem creation, or
does this happen automatically?
(In any case, having these 2 mechanisms which don't know
of each other seems wrong. Shouldn't the bad144 on-disk info be
used to initialize the "dkctl" tables, and to save it
permanently?)
best regards
Matthias