woods@most.weird.com (Greg A. Woods) writes:

> [ On , September 21, 1998 at 23:33:17 (-0400), Paul F. Wells wrote: ]
> > Subject: Re: Is it really the disk drive or something else?
> >
> > The drive is the only scsi device and it's internal; the only jumper
> > I had to fiddle was the motor start so it would spin up on the 386 boxes.
> > I replaced it when I put the drive back into the sparc.  It's been a while
> > since the drive has been outside of the sparc and I don't recall its
> > termination state.  I can check again.
> 
> My recollection is that the internal drives should be terminated.  I've
> got all my disks in an external enclosure.  Surprisingly I'm getting by
> with just a normal passive terminator, and I've got probably the maximum
> length of cable on the bus too (3m).
> 
> Internal drives must be connected by a cable no more than 30cm and no
> less than 10cm in length, with the 10cm length being preferred.
> 
> If you have all original Sun hardware then you should be OK on that
> front, provided it's in good condition.
> 
> Do check the termination....
> 
The drive must originally have been unterminated.  When I tried to terminate it,
passive/active, the prom would indicate an I/O error and not boot.  So I left it
unterminated.

> Speaking of motor start though....  You might want to jumper the drive
> so that it spins up on power-up.  I've got one drive in my system that
> seems to be set to spin up only on demand, and though the initial SCSI
> bus reset and probe seems to spin it up, I'm not 100% certain all
> versions of SS2 firmware will do this successfully.
> 
This prom is 2.4.1 and does cause the drive to spin up.  I notice, however,
that after several power cycles the drive is making different clickety/clack
noises as it spins up, so I'm going to leave it so that it spins up on
power on.

> If you have some way of connecting the drive externally you might try
> that too.  This will eliminate stress on the internal power supply,
> which may also be a factor.  (note that the external bus must definitely
> be terminated if there's anything on it)
> 
> > Sure, in fact, I should have tried to load windows or something on the
> > 'failing' disk but didn't...I can try that later.
> 
> That would be a half decent test, I suppose.
> 
> How about loading NetBSD/i386 on it instead!  Even booting NetBSD/i386
> from floppy and playing around on the drive (create partitions and
> filesystems, reboot and cycle power, then copy files from the floppy to
> the disk, etc.) should be a good test.
> 

After all that messing around with jumpers, it's working correctly, with
its original settings.  I'm guessing now that either something may have 
worked itself loose or dirty/corroded.  Or, now that I have the case
uncovered, it is just a little bit cooler...but I have this sinking feeling
that there's something expensive that's broken.


Thanks for all your help,
Paul
-- 
Paul F. Wells                      <paul@wellserv.com>