Networker

[Networker] mixed media problem or: putting an optical into a DLT?

2003-07-17 04:45:49
Subject: [Networker] mixed media problem or: putting an optical into a DLT?
From: Gaddy <xy.0815 AT GMX DOT NET>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Thu, 17 Jul 2003 04:45:42 -0400
Hi Networkers,

as new generations of drives arrive
(i.e.: AIT-1,2,3,SAIT   LTO1,LTO2, 9940A,9940 ...)
tape library manufactures allow to upgrade their libraries
with the new technology. Pretty good.

But NetWorker doesn't allow to do it
(O.K. you can use AlphaStor - but I don't think
that's the right solution for the problem).

We recently added some new STK9940B to a STK9310,
which has half a dozen of STK9840A already.

Unfortunately NetWorker tries to mount
a)9940A volumes in 9940A drives
b)9940A volumes in 9980B drives
c)9980B volumes in 9940A drives  and finaly
d)9980B volumes in 9980B drives.

No question: a+d is fine, but b+c seems to be wrong.
Playing with pools isn't the fix, at least when it comes
to restores.

So I opened a discussion with our NetWorker support
here in Germany and I'd like to here your comments:

---------------------------------------------------

Pls. let me make some remarks to the ?mixed media? problem described in
call #20289 before you are going to close the call.

1st we have to get around the ?problem? (pls. note that I don?t use
>>unsuported configuration<<) and the needs
2nd I?d like to discuss the term ?unsuported configuration? a little bit,
because the conclusions from that discussion will affect the solutions we
both can find for the ?problem?

So pls. allow me to start with item #2:

Mixed media is ?in no way? a problem of pools (altough we are not going to
use mixed media pools here).
Mixed media is only a problem of the physical configuration, because we
have well defined drive types and well defined media types.
The rule is simply: ?Don?t put diesel fuel in an unleaded engine!?. It?s
that simple, but Legato doesn?t care.
As I mentioned before, all the information is there: ?well defined device
types? AND ?well defined media types?. And it seems to be pretty easy to
check, that these two types match and not to mount an optical in a Magstar
drive. (With a little pressure - I?ll bet - you can do that ? and Legato
will do it, as described in TB371).

Pools are only a logical separation of data streams and only a <not realy
functioning> workaround to separate physical devies.

So from my point of view, not to check if device types match, is nothing
more than a serious BUG. (But it?s much easier to declare
it ?unsupported? ? which is <unfortunately> often the way of getting
around bugs).
But ?jbconfig? still asks (at least in V6 und earlier) for the device type
of every single drive in a jukebox, altough it?s not supported? Funny!

So I now come to item #1:
a) Do you expect to get the ?bug? fixed when opening a call with
Legato?  (I supose the answer is ?No?.)
b) Does it make sense to open a RFE for that ?problem?? How long does
it usally take?
c) In the short time can we do the following (???):
- Define\license a second silo on the same ACSLS and same physical
STK9310 robot.
- Define all the new 9940B drives in that jukebox (and only in that)
- Define all the 9940B volumes (not the slots ? because it?s a silo)
in that jukebox (and only in that)

Do you think this will solve the problem and not open new ones? (i.e. same
physical silo defined twice, ?)

-------------------------------------------------------------


Comments, tips, hints welcome!

Thanks -sg-

--
S.Gattert; VisoPlant GmbH - Hamburg

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

<Prev in Thread] Current Thread [Next in Thread>