ADSM-L

mixing 3590J & 3590K's in a 3494 for ADSM

2000-09-28 21:28:42
Subject: mixing 3590J & 3590K's in a 3494 for ADSM
From: Joe Faracchio <brother AT SOCRATES.BERKELEY DOT EDU>
Date: Thu, 28 Sep 2000 18:27:36 -0700
Can you say "Almost pregnant."

Well I thought I was having limited success and started composing this
letter when I noticed "anr1405w no scratch available" and realized I
screwed up my system.  But no worry.  I deleted my drive and libr defs and
redefined them and then checkin'd everything and my only problem now is
finding the scratches that got checked in as private :-)

But read on... maybe if I define separate scratch and private pools ????
... joe.f.
P.S. kids: don't try this at home unless you have a strong stomach.
 =======================================================================

I decided to try some things today to see if J's and K's can co-exist in
the same library for ADSM.  My intention is to have J's for offsite and
K's for onsite until such time as we need more offsites and J's are all
used.  (Being a 'small pc' backup service I try and keep my collocated
tapes at 50% filling and having K's for that would help.)

I don't have any K's yet and need to get the upgrade to my E drives to
handle both J's and K's.  But it appears I  have simulated two seperate
pools and got away with it. (almost, I think.)

I am currently running 3.1.2.20 on an NSM/3466.  Soon to go to 3.7.2

1. define second Library with same private category but a 'new' scratch
category,  that's pointing to original library:  (Name changed only from
LIB to L2B.)

def libr 3494L2b libt=349x device=/dev/lmcp0 privatecat=400
 scratchcat=402   ( 401 is the original scratch category)

That 'took' but it complained that all my tapes in the original
library: "3494LIB" had the same private category ... (?) with the new
library:

ANR8774W Volume 200042 not checked into library 3494L2B
 but is using category 400.

(don't know how 'bad' that is??!!!??!)

I defined a new tape devc of 3590D (D for double-length) and pointed it to
the 'new' library:
(I already separated my onsite and offsite devc's by having 3590 and
3590C defined.)

def dev 3590D devt=3590 libr=3494l2b
ANR2203I Device class 3590D defined.

I then defined a 'new' drive that co-exists with current drive.
(I always wondered why you do the name and dev seperate.)
So I defined /dev/rmt5 as also being mt6 as well as mt5.

define drive  3494l2b mt6 device=/dev/rmt5
ANR8404I Drive MT6 defined in library 3494L2B.

(BTW the above needs to be done when drive not in use.)

Checked a tape out of the current 3494lib and into 3494l2b:

checkin libvol  3494l2b 200113 checkl=no status=scratch devt=3590

ANR8422I CHECKIN LIBVOLUME: Operation for library 3494L2B
 started as process 409.
ANR8319I 024: Insert  volume 200113 R/W into library
 3494L2B within 60 minute(s).
ANR8427I CHECKIN LIBVOLUME for volume 200113 in library
 3494L2B completed successfully.

then I ran a db backup with devc=3590D and it used 200113 that I had
previously check ed into this 'other' library.  And it ran.

I also check'd in an already used db backup to the new library as scratch
and it accepted it, so that's a problem.  ADSM might write over it?

I would think that if I have a different number series of 300,000
alongside the 200,000 series I'm using and was careful to checkin tapes
only to the appropriate, library then J's and K's could co-exist.

So the only question in my mind is:  do I need to have two private
categories?  Is the above error messages about same private category
serious?  What happens if an operator checks in a private (in-use) tape to
the wrong library as scratch?  Does adsm reject it on the first attempt of
use or write over it?
Are there any other questions??
 =======================================================================
Are there other questions!!  Yes!  Why does my scratch pool get
lost??? :-)   and does this mean that defining a scratch pool
always overrides  the current one as happend to me or is there another
way of doing it so that it works?

Do I need to re-customize my 3494 to accept two scratch pools, etc?

cheers ... joe.f.



Joseph A Faracchio,  Systems Programmer
<Prev in Thread] Current Thread [Next in Thread>