[SunRescue] So I'm trying to get this Axil 220/245 "up", and I can't...

[Greg A. Woods]

[ On Monday, April 16, 2001 at 17:35:49 (-0400), Ken Hansen wrote: ]
> Subject: [SunRescue] RE: rescue digest, Vol 1 #1225 - 14 msgs

[[ I really don't like it when a thread changes subject to that of a
digest posting.  I generaly delete replies to digest postings without
ever looking at them.  It's not that my mailer threads only by subject
header, but that's more or less how I decide which messages to read, and
if the subject header is meaningless I generally don't read the message
to find out anything more.  It would be really nice if people who
receive the list as a digest could remember to adjust the subject of any
replies they might send to match the original thread, and of course if
you want *me* to even consider reading them then that's pretty much
necessary.  I only read this one because I don't remember Ken having
posted to any other threads recently.  I've corrected the subject in
this reply.  ]]

> Well, Probe-scsi shows the drives fine, as expected (and I am not
> going to type allthat info anyway ;^)

(I thought maybe you had the console on the serial port so could just
cut&paste from whatever emulator you might have had....)

OK then, physically the bus sounds OK.  "probe-scsi" can often work OK
on an improperly terminated bus because the data packets it sends and
generates are short and bursty.

> Internal Termination sounds like a stretch to me, but I'll take a good
> look into it.

Internal termination is just as important as external termination.  In
fact it's one in the same thing.  A SCSI bus "MUST" be terminated in
exactly two places:  each end (and absolutely nowhere else).  Usually
(see the note about older sparcstations at the end) the host adapter is
considered to be in the middle of the bus, which means the internal
*and* external sides of the bus must both end with terminators.

> The Axil 245 is a lunchbox (sorta) sized machine, with the PS/drives
> infront of the MB, not over it (like in say, an LX) so the foot print
> is a bit bigger.

Is there an internal SCSI bus (i.e. a single cable with multiple
connectors), or are there multiple connectors on the motherboard with
one stub cable for every device (as in an older sparcstation)?

If it's a bus cable is there a small black package (i.e. not a
connector) at the end?  If so then that's the terminator pack for the
internal half of the bus.

> The SCSI bus works fine w/o an external CD, but have not tried an
> extenral HD - I will.

Many SCSI CD's (except maybe Sun branded ones) have termination enabled
by default.....

> The internal drive may not be terminated properly, but I don't think
> that is the case, since I was able to install on to this drive in
> place as is. (If I swapped this drive with a pre-installed OS from
> another machine I would consider this a more likely cause)

There's only one situation where an internal drive should be terminated
-- that's when the internal cable is a bus and when there is no
terminator at the end of the cable and when the drive is connected on
the end of the cable.  Note those are all "and"s....

You really should check that the bus termination is exactly correct,
which means checking that the internal drive either has its termination
enabled or disabled as is appropriate.

Note also that if the internal devices are on stub connectors, as they
are in any older sparcstation (eg. 1, 2, 1+, etc.) then the host adapter
is considered to be at the end of the bus and is itself terminated, with
the internal connectors being "stub" connectors which must not be
terminated.  This also means that no external termination is required if
there's no cable plugged into the external connector because the onboard
wiring between the host adapter driver chip(s) and the external
connector are also effectively of stub length.  There may also be some
form of auto-termination too (I've never seen the actual schematics).

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods at acm.org>     <woods at robohack.ca>
Planix, Inc. <woods at planix.com>;   Secrets of the Weird <woods at weird.com>