I've never had a problem defining multiple TSM (logical) libraries on one
device address (but I can't say I've tried it since 6.2, and that was on
Windows).
What you can't do is have one device class pointing to 2 different libraries,
so you'll also have to do some juggling there, create some new devclasses and
storage pools to use going forward.
Wanda Prather
TSM Consultant
ICF International Cybersecurity Division
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Skylar Thompson
Sent: Thursday, December 11, 2014 10:15 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] Drive preference in a mixed-media library sharing
environment
Interesting. I hadn't considered using different libraries to solve this.
It was a little unclear from the thread - does this require partitioning on the
library side? I wasn't aware that two different libraries (presumably with two
different paths) could share a single device special node.
On Wed, Dec 10, 2014 at 06:23:10PM -0600, Roger Deschner wrote:
> It won't work. I tried and failed in a StorageTek SL500 library with
> LTO4 and LTO5. Just like you are reporting, the LTO4 tapes would get
> mounted in the LTO5 drives first, and then there was no free drive in
> which to mount a LTO5 tape. I tried all kinds of tricks to make it
> work, but it did not work.
>
> Furthermore, despite claims of compatibility, I found that there was a
> much higher media error rate when using LTO4 tapes in LTO5 drives,
> compared to using the same LTO4 tapes in LTO4 drives. These were HP
> drives.
>
> The only way around it is to define two libraries in TSM, one
> consisting of the LTO5 drives and tapes, and the other consisting of
> the LTO6 drives and tapes. Hopefully your LTO5 and LTO6 tapes can be
> identified by unique sequences of volsers, e.g. L50001 versus L60001,
> which will greatly simplify TSM CHECKIN commands, because then you can
> use ranges instead of specifying lists of individual volsers. To check
> tapes into that mixed-media library I use something like
> VOLRANGE=L50000,L59999 on the CHECKIN and LABEL commands to make sure
> the right tapes get checked into the right TSM Library. Fortunately
> the different generations of tape cartridges are different colors.
>
> You can read all about what I went through, and the good, helpful
> recommendations from others on this list, by searching the ADSM-L
> archives for "UN-mixing LTO-4 and LTO-5". Thanks again to Remco Post
> and Wanda Prather for their help back then in 2012!
>
> Roger Deschner University of Illinois at Chicago rogerd AT uic DOT
> edu
> ======I have not lost my mind -- it is backed up on tape
> somewhere.=====
>
>
> On Wed, 10 Dec 2014, Grant Street wrote:
>
> >On 10/12/14 02:40, Skylar Thompson wrote:
> >> Hi folks,
> >>
> >> We have two TSM 6.3.4.300 servers connected to a STK SL3000 with 8x
> >> LTO5 drives, and 8x LTO6 drives. One of the TSM servers is the
> >> library manager, and the other is a client. I'm seeing odd behavior
> >> when the client requests mounts from the server. My understanding
> >> is that a mount request for a volume will be placed preferentially
> >> in the least-capable drive for that volume; that is, a LTO5 volume
> >> mounted for write will be placed in a LTO5 drive if it's available,
> >> and in a LTO6 drive if no LTO5 drives are available.
> >>
> >> What I'm seeing is that LTO5 volumes are ending up in LTO6 drives
> >> first, even with no LTO5 drives in use at all. I've verified that
> >> all the LTO5 drives and paths are online for both servers.
> >>
> >> I haven't played with MOUNTLIMIT yet, but I don't think it'll do
> >> any good since I think that still depends on the mounts ending up
> >> in the least-capable drives first.
> >>
> >> Any thoughts?
> >>
> >> --
> >> -- Skylar Thompson (skylar2 AT u.washington DOT edu)
> >> -- Genome Sciences Department, System Administrator
> >> -- Foege Building S046, (206)-685-7354
> >> -- University of Washington School of Medicine
> >might be a stab in the dark ..... try numbering the drives such that
> >the LTO5's are first in the drive list or vice versa.
> >That way when tsm "scans" for an available drive it will always try
> >the LTO5's first.
> >
> >HTH
> >
> >Grant
> >
--
-- Skylar Thompson (skylar2 AT u.washington DOT edu)
-- Genome Sciences Department, System Administrator
-- Foege Building S046, (206)-685-7354
-- University of Washington School of Medicine
|