Debugging SCSI ... why so hard?

[Tom Samplonius]

On Thu, 24 Apr 2003, Marc G. Fournier wrote:

> I'm just curious about this ... with how 'intelligent' SCSI is (isn't
> it?), and with the newer standards, why is it so difficult to isolate
> issues with SCSI?  For instance, cabling problems ... isn't there any way
> of detecting, based on signal (strength or something) whether there is a
> termination problem?

  There is just so much involved with SCSI.  Unlike, IDE/ATA, SCSI can be
used to connect more than two devices, so it requires termination. Plus,
so the SCSI bus is longer and has more connectors, each of which can
induce problems.  I've never had any problems that weren't caused by using
sub-standard components.  And when things aren't right, you get lots and
lots of erros!  Even with four drives on a bus, using a process of
elimination, it doesn't take long to figure out what is wrong.  

  SCSI backplanes really are the way to go, though.  There is just a
single cable between the SCSI controllor and the backplane, and the
backplane has a built-in terminator. And the drives connect directly to
the backplane, and their SCSI id is set by their position in the
backplane.

Tom