Craig,
First of all, I would like to thank you very much for asking the user
community about this before making the change. This type of advance
communication is very much appreciated! To answer your questions:
I have never needed to set the mountlimit for a device class to anything
other than the number of drives currently available in that class. Indeed,
I always found it a nuisance to have to both delete a drive and also
decrement the mount limit for the device class.
While you are working in this area, I have often wished for the ability
to temporarily quiesce, or vary offline, one of the drives in a library,
so that I could clean the drive, or perform some other activity on it.
As things work now, I must DELETE the drive, decrement the mount limit,
clean the drive, then re-DEFINE the drive, and increment the mount limit.
This process is error-prone, especially when I forget to write down what
the definition of the drive should be (e.g., especially the element number).
We will be looking into using NetTape to share drives between ADSM and
other servers. We don't know much about it yet, but I wonder if there
might be some reason to have a lower mount limit in such an environment???
I have to confess that I don't yet understand fully how ADSM works with
NetTape. (we're planning to do this on an SP2) I don't understand how
drives are allocated to ADSM when using NetTape.
You ask about having multiple device classes in one library. While we do
not currently do this, I anticipate that we will want to do this in our
Q47 DLT library. We will probably want to add newer, higer capacity
drives to our library which will leave us with two different drive types
(and therefore device classes) in the same library, but both would have a
device type of DLT. This is very similar in concept to having both 3490
and 3590 drives in the same robot, but in this case they would share the
same device type. I could see the same sort of problem arising in a 4mm
or an 8mm library, which had 2 different types of drives in them.
Paul Zarnowski Phone: 607/255-4757
Cornell Information Technologies Fax: 607/255-6523
Cornell University US Mail: 315 CCC, Ithaca, NY 14853-2601
--
On Thu, 17 Oct 1996 14:21:17 PDT BELL, CRAIG said:
On Thu, 17 Oct 1996 14:21:17 PDT BELL, CRAIG said:
>Have you ever been bitten by deleting a drive from adsm and not
>decrementing the mountlimit? ADSM assumes it has more drives available
>than are defined to the library and many transactions fail with the
>message
>
> 8447E No drives are currently available in library
>
>I am working a change to the ADSM server (for unix's, OS/2, and NT
>platforms) that will better tie the MOUNTLIMIT of a device class to the
>number of drives in the devclass's library. Does anyone see a problem
>with this? There's no reason in my mind that the devclass's mountlimit
>should ever be set to anything other than the number of drives of that
>Device Type that are defined in the library. But do any customers have
>more drives defined to adsm and perhaps set the mount limit lower in
>order to control drive usage? Also are there customers that have
>multiple device classes (of the same device type) that use the same
>library? In this case, it is the total of ALL the devclass's
>mountlimits that must be set to the number of drives. If you have either
>such a configuration, I woul really like to know what the reason is and
>to see if there is another way to accomplish your purpose. Otherwise,
>this may be restricted in a future version.
>
>I'm not talking about 3494 libraries that contian both 3490 and 3590
>drives- these are different device types. So are 3995 optical jukeboxes
>that have both WORM and OPTICAL device clases defined to them.
>
>Craig Bell
|
