Subject: Re: State of the Universe (Was: Re: SPARCstation 1 and Archive Viper 150)
To: None <port-sparc@NetBSD.ORG>
From: Ken Hornstein <kenh@cmf.nrl.navy.mil>
List: port-sparc
Date: 02/12/1996 01:56:07
>Anyway, in looking at this situation, I asked myself what it is that I depend
>on most in my at-home environment.  Basically it's
>
>	(1) my Exabyte 8500, since at-home backups are an absolute *necessity*

It's my understanding that different Exabytes have different firmware which
tend to vary in their friendliness to non-Sun systems.

My personal experience with tape drives are as follows.  Note that this
is all under NetBSD/i386, but since everyone now uses the MI SCSI code
(err, that _is_ correct, right?), I think this should apply to everything.

- I used an Archive viper under 0.9-current, which "mostly worked".

- More recently, I've had the occasion to try one of the newer HP DAT drives,
  an Exabyte drive, and a Sony DAT stacker.

  The HP totally did not work.  Writes to it would start to generate "command
  aborted" sense messages (which is why I sent in a PR about that not too
  long ago :-) ), but after a while it just sent out a continuous stream
  of "command aborted" to all writes.  Reads were worse - it read for a
  little bit, then totally hung (I have since discovered there are a bunch
  of jumpers on it for making it work on i386 hardware, but I haven't had a
  chance to try the drive again).  Works like a charm on Solaris and
  HP systems.  I suspect that the Solaris and HP SCSI systems are just
  smarter, or use a different way of talking to that drive, because there
  was some seriously weird stuff going on with it.

  The Exabyte worked just fine, once I turned off adapter sync negotiation
  on my SCSI card.  Bugger slow as all hell, though.

  The Sony DAT worked right out of the box, no problems.

So I wouldn't call tape support "non-existant".  I think right now it's more
of a "try it and see" thing.  But some stuff definately does work.

--Ken