[Sussex] Diagnostic boot disk

[Ronan Chilvers]

On Sat, 2 Jul 2005 23:39:48 +0100
Steve Dobson <steve at dobson.org> wrote:

> To be honest the sort of simple checks you're asking for are mostly
> done by the BIOS before the system boots.  If the BIOS finds fault
> with the memory or video card then is make a sound (you better hope
> the speaker works).

Yes, but the difference is I need a non-techie to be able to see and
act upon these checks.

<snip>
> it locks up that part of the system.  The reset of the laptop is
just
> fine, but the CD IO thread is blocked (I assume waiting for an
> interrupt that is never going to happen, or something like that).

So that would be discovered by running a read test on the drive?

<snip>
> just use a live CD and if the device works with that it must be
> okay.  

True, but much too slow.

> BTW - I have just burnt 28 copies of Knoppix 3.9 for the BCF tomorrow.
> Stop by and buy one.  :-)

Was picking strawberries, onions and carrots at my local PYO... linux
CD or food... close call!!

Maybe if I outline exactly what its for it might clarify what I'm
asking.  I do some bits and pieces for a charity over in Hailsham in
East Sussex (http://www.computersforcharities.org).  We refurbish
donated PCs (and Macs actually), server hardware, UPSs, etc, etc and
ship it out to school projects in Africa. We get through large numbers
of PCs and our throughput is increasing - currently we refurbish c.100
machines a month (that's 25 machines every Tuesday evening - we can
only get together once a week at the moment), but we need to step it up
to about 250 to 300 to cope with increasing demand. For example we are
currently working on 2 projects, each requiring 200 machines, one for
Zimbabwe and one for South Africa.

Currently the refurbishment consists of checking over the hardware,
fixing simple problems (such as listed in previous email - RAM/CDROM/
FDD/HDD/etc), ie: that can be solved by swapping out bits, then zeroing
the HDD and cloning a pre-prepared W2K image onto it from a networked
server.  We don't go into depth with fault diagnosis, firstly because we
haven't got time and secondly, we have sufficient influx of hardware to
cope with _some_ losses.

The main problem we have is technically able bodies (volunteers
anyone?).  There are only really 2 or 3 of us who are able to diagnose
and fix hardware problems.  We do have several other volunteers though
who are up to following a step-by-step cloning howto, or checking that
monitors are working, etc, etc.  We need to streamline or spread the
fault diagnosis bit so that when it gets to us techies, we have a
reasonable idea of what needs to be done.  Hence if I had a handy magic
boot disk that a non-techie person could stick in the FDD, test the
machine and then write a sticker which said, 'CDROM not reading' or
'RAM bad', that would speed up the repair process hugely.  There are of
course, as you say, a huge number of possible hardware combinations and
we do indeed see massive variations in configuration.  But we do also
see large batches of similar machines donated by, for example, colleges
or business who have 300 Compaq T450s or Dell GX110s they want shot
of. All basically similar hardware, with maybe HDD size or RAM size
variations.

Sorry about the length of that but I hope it clarifies the requirement
a little more.  A live CD takes far too long to be usable.  We need a
diagnosis that lets a non-techie make a good stab at general fault
finding within, say, 5 minutes (from boot).  They might need to
get through 30 to 40 machines (or more) in 3 to 4 hours.

Any ideas appreciated.

Cheers

Ronan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mailman.lug.org.uk/pipermail/sussex/attachments/20050703/81a7a991/attachment.pgp