disk scrubber for Linux?

Paul Lussier p.lussier at comcast.net
Fri Dec 9 09:01:00 EST 2005


"Michael ODonnell" <michael.odonnell at comcast.net> writes:

> Otherwise, please allow me to continue to at least suggest the
> laughable possibility that periodic reads of disk blocks to
> verify their integrity isn't a total waste of time, and that
> we're still interested in hearing about any apps that might
> implement this approach to enhancing RAID reliability...

As I recover from laughing hysterically at this reply, and force
myself not to report this post to the Center For Over-Reaction, please
allow me to joing in mod's request, as I too am quite interested in
such a beast.

I'd also like to add that SMART, though often helpful, is more often a
daemon who's job it is to inform you that it's just shut the barn door
now that the palomino is long gone.

We've been seeing some incredibly high drive failure rates among
Maxtor drives, and I believe Hitatchi has just joined the list with a
failure similar to what was also described here.

A while back I found this url:

  http://smartmontools.sourceforge.net/BadBlockHowTo.txt

Which does a very good job of describing an unfortunately tedious,
manual, and very difficult to automate, method of identifying a file
associated with an unreadable disk sector, and how to force that
sector to reallocate.  We've found SMART to be a decent indicator that
a drive needs block-reallocation, but no decent method of dealing with
the situation other than re-formatting the system's drive and
re-installing (thank you Debian and FAI :)

This is fine for our test/development systems.  Not so much for
customers with installed systems where this would result in data loss.

-- 

seeya,
Paul



More information about the gnhlug-discuss mailing list