Path: utzoo!attcan!uunet!know!zaphod.mps.ohio-state.edu!maverick.ksu.ksu.edu!rutgers!mcnc!rti!bcw
From: bcw@rti.rti.org (Bruce Wright)
Newsgroups: comp.os.msdos.apps
Subject: Re: SPINRITE
Summary: Spinrite & its usefulness
Keywords: Does it really work?
Message-ID: <1990Nov18.201156.16054@rti.rti.org>
Date: 18 Nov 90 20:11:56 GMT
References: <666@seer.UUCP> <Pc9Js3w163w@kk4fs.UUCP>
Organization: Research Triangle Institute, RTP, NC
Lines: 51

In article <Pc9Js3w163w@kk4fs.UUCP>, ambush@kk4fs.UUCP (Ambush Bug) writes:
> My experience with Spinrite has been:
> Yes, it reformats your hard disk, but it's so slow that unless you've got a 
> 330 meg job you're backing up onto 360K floppies, it's faster to back it up 
> and then reformat it.
> Yes, it can recover bad sectors, but they'll go bad again as soon as you 
> write something to them (adios, data!)

I disagree.  It's true that Spinrite will take longer than backing up
and restoring the disk, but it saves having to feed floppies the whole
time.  You can run it overnight and have the disk reformatted in the
morning;  unless you happen to have a tape drive that is > the size of
your hard drive, you can't do a complete disk restore without having to 
be on hand to feed floppies or tapes to the system.

You do have to be careful about doing this if there's critical data -
you ought to back up the disk first.  But you ought to keep backups
reasonably current anyway, Spinrite or no;  then the incremental cost
of running Spinrite would be an incremental backup which should be no
big deal if you are keeping current :-).

In my experience, the only time that sectors go bad again is when the
drive is on its way out, though it's possible that in some unusual
cases a few bad sectors may pass all of the pattern tests (pattern
testing is probabalistic: a few "worst-case" patterns get most of
the bad sectors by testing things like alternating patterns, all-0's,
all-1's, etc;  but adding more patterns will always find a few more
sorts of things that could be wrong with the surface).  That's why
you should always run as many pattern tests as possible (and why
Spinrite recommends that you not return previously-marked bad sectors
unless you run their most extensive pattern test).  Many manufacturers
low-level formatting doesn't do any more pattern testing than Spinrite,
so you're not gaining much (if anything) by using their test instead
(in my experience).  You can always run Spinrite and NOT return the
bad sectors;  this can actually be useful in some cases where the
drive has "soft" (recoverable) errors which are usually cleared up
by this procedure.

> All in all, it's not worth it. It's slow and clunky. What's more, you can 
> really screw things up with it.

This last is true of a lot of things - notably the Norton Utilities
(and similar sorts of programs).  I think it's a little like saying
that you can really mess things up by using something like a circular
saw:  it can cut quickly and can be dangerous to life and limb (and to
your project) if used improperly.  But I don't think that that's an
argument not to use it;  I think it's an argument to use it _properly_,
since all of these things can be quite useful tools if treated with
respect.

						Bruce C. Wright