SUMMARY: Networker & HP-C1553A

From: Stefan Rauthenberg (rauthenberg@hhi.de)
Date: Mon Sep 04 1995 - 02:20:24 CDT


Hello,
About a week ago I asked the following question:

---
I am trying to install a HP autoloader HP-C1553A for Networker 4.0.2 
with the 6 slot jukebox option.  

After installing and configuring the jukebox driver a simple test is recommended to be run which failed as follows:

Note: the outcome seems to be independent whether the networker patch [101955-02] has been installed or not.

# jbexercise -c /dev/sjid1u1 -m HP-C1553A About to exercise the jukebox. There should be media loaded in the jukebox in the first and last slots. Ready to continue(y/n)?y slot 1: "FULL" slot 2: "EMPTY" slot 3: "EMPTY" slot 4: "EMPTY" slot 5: "EMPTY" slot 6: "FULL" jbexercise: Jukebox error, HP DAT AutoChanger not ready, operator intervention may be required: Sense Key <2>, ASC <4>, ASCQ <1> (45025) I wonder if anybody managed to get it work. ------

To make browsing of summaries easier I'd like to begin with the software version and the networker host experiencing the problem:

Solaris Networker Advanced 4.0.2b + 6 slot jukebox option installed on a SPARCstation 10 running Solaris 2.3

What I got from the list ------------------------ I received five replies and the overwhelming majority confirmed having problems with the HP-C1553A autoloader. One person (Chris Zanelli <chrisz@ind.tansu.com.au>) recommends to use DLTs instead of the HP dat stacker due to his bad experiences with such a device. Another common advice was to upgrade to Legato's newest version which seems to be something like 4.1.3. However, this was a no-no for me as I purchased the soft just a month ago from Sun. Another interesting and promising solution was offered by Yossi Goltz as he managed to find a way to workaround timeouts which happen to occur while networker is trying to dismount a full tape and to mount a fresh one.

Yossi Goltz <yossi@romeo.ddddf.com> wrote: ----- Begin Included Message -----

I run networker 4.0.2 (on sparc sunos 4.1.3) with the HP autoloader. I found out (after a long time) that the tape and the driver do not match on one simple thing: tape unload! I've contacted HP, Legato and the vendor who sold me the tape, got a few updates for the tape drive software but nothing helped.

Finaly just before I gave up, I tried rewinding the tape before issuing the nsrjb command to unload the cardtridge. this worked.

I found out that the drive when asked to unload a cartridge starts off by a rewind operation and only when it's on the begining of the tape it unloads. Since I use 120 meter tapes this cantake quite some time. On this time the driver times out!

If the tape is close to the beggining it works ok but if it's at the end it will hang the driver.

you can go round the problem if you change the load command networker uses to load the next tape to a rewind command followed by the original nsrjb command. I did this and it works for me.

----- End Included Message -----

The timeout problem hit me after I managed to run `jb_exercise' successfully but a change of the tape-config-list was the adopted solution.

Solutions: --------- The errors during the run of `jb_exercise' disappeared after I changed the tape drive's switches to:

sw 1 2 3 4 5 6 7 8 1 1 0 1 1 1 0 0

The correct switch settings were provided by Legato's speedy technical hotline (support@Legato.COM)

Although `jb_exercise' didn't complain anymore and I was able to label tapes, spurious timeouts occured and the tape drive got hung as networker was trying to switch from a full tape to the next empty volume during a nightly backup session. A typical entry in /var/adm/messages looks like:

,400000/esp@f,800000 (esp0): Aug 29 03:18:25 bssun1 unix: Disconnected command timeout for Target 4.0 Aug 29 03:18:30 bssun1 unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f ,400000/esp@f,800000/st@4,0 (st4): Aug 29 03:18:30 bssun1 unix: SCSI transport failed: reason 'timeout' Aug 29 03:18:31 bssun1 unix: : retrying command

Again I contacted Legato's hotline and received a tape-config-list for Solaris 2.3 which is different from the one given in the release notes packaged with Solaris Networker 4.0.2.

tape-config-list = "HP C1553A ", "HP C1553A DAT Loader", "HP-DAT"; HP-DAT = 1,0x34,1024,0xd639,1,0x00,0;

N.B. This is for Solaris 2.3 and should not be used for 2.4, since the entry <bsize> is just a bug fix for 2.3 and its meaning is being overridden by the flag ST_VARIABLE. A value of 0 should be ok for 2.4.

After installing the new tape-config-list, the timeouts disappeared and I've been using the stacker without any problems.

Support Issues -------------- I purchased the networker product from Sun and learned that I am supposed to contact Sun and NOT Legato in case of technical problems. As a matter of fact I tried Sun first but the only thing they managed to come up with was two non-working switch settings. Since the problem wasn't solved after a week, I tried Legato as a last resort. They responded within hours taking into account the time lag and most important the solutions offered hit the nail on the head.

Lesson learned:

"Buy networker products from Legato or a source which entitles you to receive support directly from Legato."

[I encourage Sun to prove me wrong]

Many thanks to the following persons for their valuable responses:

bharel@digit.com (Bruce L. Harrell) Yossi Goltz <yossi@romeo.ddddf.com> Chris Zanelli <chrisz@ind.tansu.com.au> "Brian J. Doherty" <bdoherty@odd.Stanford.EDU> juntunen@ocelot.trs.ntc.nokia.com (Timo Juntunen)

Stefan ________________________________________________________________________ Stefan Rauthenberg Heinrich-Hertz-Institut fuer Nachrichtentechnik Berlin GmbH Phone: +49 30 31002-615 Einsteinufer 37 Fax: +49 30 3927200 D-10587 Berlin - Germany Email: rauthenberg@HHI.DE ________________________________________________________________________



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:32 CDT