> 1) this wreaks havoc with scsibusses with multiple initiators on the bus
> (vulgo: HA clusters).

Yep - HA cluster definitely do not like, at all, one, bit, not at all,
having the other initiator reset the scisi bus.  Someone I know spent
several weeks in NZ (from the UK) because of it.

Not the least of the problems is that you start finding bugs in the disk
software - so only certain revisions of certain disk from specific
manufacturers work properly.

For full effect, use a sci <=> fddi <=> scsi bridge set in order to get
the second set of disks far enough away from the primary set that they
are likely to be unaffected by a major catastophe (eg an aircraft hitting
your computer center - the AA [1] built a second building many years ago to
ensure their service would not be affected by such an event).

	David

[1] UK motoring organisation

-- 
David Laight: david@l8s.co.uk