[comp.sys.apollo] Exabytes on Series 400's

lampi@polari.UUCP (Michael Lampi) (06/22/91)

I've been testing Exabyte 8200's and 8500's on various HP/Apollo 9000
Series 4XX's (400t, 425t, etc.) under Domain/OS and (sigh) HP-UX, and have
a couple of observations, and a couple of questions.

First, 8200's work under HP-UX in all cases. No ifs, ands or buts.

Second, 8500's work under HP-UX in all cases IF they have a brand-new firmware
upgrade that just got released last week. Again, no changes required to the
OS and no 3rd-party drivers required.

Third, 8200's and 8500's work under Domain/OS in many cases, but not on a
Mentor node, and tar sometimes has i/o troubles that are bit-pattern specific
that don't occur with wbak or rbak. 

"Not on a Mentor node" means that the
node crashes immediately with status 440002, meaning "bad iova from os/dio
bus manager" when using rbak, wbak and rwmt, but tar, cpio, mt, etc., all
execute flawlessly, and netstat -c reports no tape drives while scsi_info
reports a tape drive at SCSI ID 1.

Has anyone else seen this behavior? Any solutions? Any reason why HP has
decided to support 8mm tape drives under HP-UX (or Unix in general) and
not support the Domain/OS utilities?

Thanks in advance.

-- 
-----
Michael Lampi            polari!lampi@sumax.seattleu.edu

krowitz@RICHTER.MIT.EDU (David Krowitz) (06/24/91)

I've got an additional Exabyte problem for you ... we've
got one of the older 8200's attached to a DN3500 running
off of the SCSI interface board provided by Workstation solutions.
I tried running the drive off the WD7000 SCSI interface,
but the node would crash nearly every time the drive was
power cycled or whenever we got even the tiniest bit of
static discharge. This is with SR10.2, not SR10.3, by
the way. We also saw these crashes on DN2500's.



 -- David Krowitz

krowitz@richter.mit.edu   (18.83.0.109)
krowitz%richter.mit.edu@eddie.mit.edu
krowitz%richter.mit.edu@mitvma.bitnet
(in order of decreasing preference)

thompson@PAN.SSEC.HONEYWELL.COM (John Thompson) (06/24/91)

> Third, 8200's and 8500's work under Domain/OS in many cases, but not on a
> Mentor node, and tar sometimes has i/o troubles that are bit-pattern specific
> that don't occur with wbak or rbak. 
> 
> "Not on a Mentor node" means that the
> node crashes immediately with status 440002, meaning "bad iova from os/dio
> bus manager" when using rbak, wbak and rwmt, but tar, cpio, mt, etc., all
> execute flawlessly, and netstat -c reports no tape drives while scsi_info
> reports a tape drive at SCSI ID 1.
> 
> Has anyone else seen this behavior? Any solutions? Any reason why HP has
> decided to support 8mm tape drives under HP-UX (or Unix in general) and
> not support the Domain/OS utilities?
At 10.1, you needed a PSK and patch (set of patches?) to get the SCSI 
devices going.  At 10.2 and beyond, these were included in the base O/S.
At 10.3, they seem to have added official support for a 3rd and 4th tape
drive (rmts8, 9, 10, 11 are all there now on our nodes, along with the
non-rewinding devices rmts12 .. rmts15).  They also added (at 10.3) 
support for the Domain utilities rbak and wbak.  The device is 'm' in this
case.  It takes a looonnnnggg time while it plays around (checking for a
real magtape device???), but then it works just fine here with a normal
10.3 load.

-- jt --
John Thompson
Honeywell, SSEC
Plymouth, MN  55441
thompson@pan.ssec.honeywell.com

Avoid the rush -- Procrastinate Now!

lampi@polari.UUCP (Michael Lampi) (06/25/91)

In article <9106241630.AA16316@pan.ssec.honeywell.com> thompson@PAN.SSEC.HONEYWELL.COM (John Thompson) writes:
->
->
->> Third, 8200's and 8500's work under Domain/OS in many cases, but not on a
->> Mentor node, and tar sometimes has i/o troubles that are bit-pattern specific
->> that don't occur with wbak or rbak. 
->> 
->> "Not on a Mentor node" means that the  [Mentor]
->> node crashes immediately with status 440002, meaning "bad iova from os/dio
->> bus manager" when using rbak, wbak and rwmt, but tar, cpio, mt, etc., all
->> execute flawlessly, and netstat -c reports no tape drives while scsi_info
->> reports a tape drive at SCSI ID 1.
->> 
->> Has anyone else seen this behavior? Any solutions? Any reason why HP has
->> decided to support 8mm tape drives under HP-UX (or Unix in general) and
->> not support the Domain/OS utilities?
->At 10.1, you needed a PSK and patch (set of patches?) to get the SCSI 
->devices going.  At 10.2 and beyond, these were included in the base O/S.
->At 10.3, they seem to have added official support for a 3rd and 4th tape
->drive (rmts8, 9, 10, 11 are all there now on our nodes, along with the
->non-rewinding devices rmts12 .. rmts15).  They also added (at 10.3) 
->support for the Domain utilities rbak and wbak.  The device is 'm' in this
->case.  It takes a looonnnnggg time while it plays around (checking for a
->real magtape device???), but then it works just fine here with a normal
->10.3 load.
->
->-- jt --
->John Thompson
->Honeywell, SSEC
->Plymouth, MN  55441
->thompson@pan.ssec.honeywell.com
->

The Mentor node claims to be at SR10.3, and it has the most problems. Other
'non-Mentor' 9000/4XX nodes at 10.3 don't seem to have these problems.

BTW, the looonnnnggg (sp?) time Exabytes spend playing around is due to the
'ANSI-like' labeling and relabeling of the tape. Writing filemarks takes the
majority of the time, and rewinding, etc. takes the rest.


-- 
-----
Michael Lampi            polari!lampi@sumax.seattleu.edu

goykhman_a@apollo.hp.com (Alex Goykhman) (06/27/91)

In article <4520@polari.UUCP> lampi@polari.UUCP (Michael Lampi) writes:
>I've been testing Exabyte 8200's and 8500's on various HP/Apollo 9000
>Series 4XX's (400t, 425t, etc.) under Domain/OS and (sigh) HP-UX, and have
>a couple of observations, and a couple of questions.
>
>First, 8200's work under HP-UX in all cases. No ifs, ands or buts.
>
>Second, 8500's work under HP-UX in all cases IF they have a brand-new firmware
>upgrade that just got released last week. Again, no changes required to the
>OS and no 3rd-party drivers required.
>
>Third, 8200's and 8500's work under Domain/OS in many cases, but not on a
>Mentor node, and tar sometimes has i/o troubles that are bit-pattern specific
>that don't occur with wbak or rbak. 
>
>"Not on a Mentor node" means that the
>node crashes immediately with status 440002, meaning "bad iova from os/dio
>bus manager" when using rbak, wbak and rwmt, but tar, cpio, mt, etc., all
>execute flawlessly, and netstat -c reports no tape drives while scsi_info
>reports a tape drive at SCSI ID 1.

    This problem was fixed in SR10.3.0.2 (patch_m0190).

>
>Has anyone else seen this behavior? Any solutions? Any reason why HP has
>decided to support 8mm tape drives under HP-UX (or Unix in general) and
>not support the Domain/OS utilities?
>
>Thanks in advance.
>
>-- 
>-----
>Michael Lampi            polari!lampi@sumax.seattleu.edu

    Alex Goykhman


-- 
Alex Goykhman                    
Chelmsford System Software Lab   mit-eddie!apollo!goykhman_a
Hewlett-Packard, Company         goykhman_a@apollo.hp.com
Needless to say, I speak for myself only