[comp.sys.mac.apps] SUM II's Achilles Heel

nick@cs.edinburgh.ac.uk (Nick Rothwell) (10/29/90)

I bought myself a copy of SUM II last week. All in all, a nice set of tools,
although the program interfaces are a little clunky at times.

Anyway: SUM II provides (as I'm sure you know) disk
optimisation/defragmentation tools, soft partitioning, and a sophisticated
file/volume recovery mechanism which traces deleted files, and can be
asked to recover entire volumes, repair directories, and so on.

It keeps track of the current system configuration (number and kind of
disks) so that it can recover devices which should be there but which
aren't mounted.

The Achilles Heel seems to be the following: to put it rather bluntly,
if something spits on the header of a soft partition file, it's f*cked.
Each soft partition is held as a single contiguous file on the main
(physical) volume. You can run the optimisation and protection utilities
on partitions as well, so you're protected from file deletion and
initialisation. But, what if the partition file becomes corrupted so as
to be unmountable? You can't recover the contents, because you can't
mount it (SUM Recover reports an error not described in the manual and,
er, refers you to the manual...). And, you can't recover the partition
file itself in the enclosing volume because it's still there, but just
corrupted.

I don't know how common this kind of failure is. It happened to me after
I was trying to optimise a partition with very little available space on
it. I presume that corruption elsewhere in the partition is just seen
as corruption of the virtual disk, and can be dealt with.

Of course, there's no substitute for taking proper backups (which I'd
done, of course), but it was rather annoying to have SUM just throw its
hands in the air and give up over a simple file corruption.

On a vaguely related matter: has anybody else noticed that the help
screens for the SUM Partition DA have junk characters in the top right
corner?
-- 
Nick Rothwell,	Laboratory for Foundations of Computer Science, Edinburgh.
		nick@lfcs.ed.ac.uk    <Atlantic Ocean>!mcsun!ukc!lfcs!nick
~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~ ~~
 "Now remember - and this is most important - you must think in Russian."

phil@waikato.ac.nz (10/30/90)

In article <1091@skye.cs.ed.ac.uk>, nick@cs.edinburgh.ac.uk (Nick Rothwell)
writes:
> The Achilles Heel seems to be the following: to put it rather bluntly,
> if something spits on the header of a soft partition file, it's f*cked.
> Each soft partition is held as a single contiguous file on the main
> (physical) volume. 
> [...] But, what if the partition file becomes corrupted so as
> to be unmountable? You can't recover the contents, because you can't
> mount it (SUM Recover reports an error not described in the manual and,
> er, refers you to the manual...). And, you can't recover the partition
> file itself in the enclosing volume because it's still there, but just
> corrupted.

Here's another:  I optimised my hard disk so that I could fit an 800K partition
on it.  Then I changed my mind and decided to go for 1 Mb.  So I defragment
again and go off for a cup of tea while it's doing it's work.  When I come back
there is an error message on the screen which told me nothing (something like
"Unexpected error") but might as well have said "I've just f**ked your data".
I suspect it was something to do with having made all the files nice and
contiguous, it dodn't want to go and chop them up again to shuffle them about
to give me my 1024K.  

It works fine if you don't try and do (admittedly silly) things like that, but
I was testing it (and I did have backups - I wouldn't test something like that
without a backup).

On the subject of testing, I've just been given Norton Utilities for the Mac to
try out.  I'll make some comments on it and comparisons between it and SUM II
if anyone's interested (in due course).
> 
> On a vaguely related matter: has anybody else noticed that the help
> screens for the SUM Partition DA have junk characters in the top right
> corner?

I just checked (it's the first time I looked) - yes - looks like what happens
when there's no character for the ASCII code - I wonder if they checked?
-- 
Phil Etheridge (phil@waikato.ac.nz)                 /\  /\   -+-,--, .--, ._
Computer Services/Mathematics & Statistics         /  \/  \  / /--< /--  /
University of Waikato, Hamilton, NZ.              /        \/ /__.) \_  /