ADSM-L

Re: Reclamation problem - tape robot gets full!

1998-03-29 18:31:07
Subject: Re: Reclamation problem - tape robot gets full!
From: Trevor Foley <Trevor.Foley AT BANKERSTRUST.COM DOT AU>
Date: Mon, 30 Mar 1998 09:31:07 +1000
Hi Rene,

I've seen this problem also. I added a stack of new scratch tapes to our
library one afternoon, and by morning, they had all been used, and then
reclaimed, so all had a status of pending. Since then we have increased
the reclaimation threshold (from 65% to 80%) and haven't seen it again.

I am assuming (hoping!!) that a fix for the original problem, will be in
pft3 when it is released. I was expecting it to be in Friday's release,
but I have only found new clients so far. Anyone know when this will be
out?


Trevor

        -----Original Message-----
        From:   Lambelet,Rene,VEVEY,FC-SIL/INF.
[SMTP:Rene.Lambelet AT NESTLE DOT COM]
        Sent:   Monday, March 30, 1998 4:14 AM
        To:     ADSM-L AT VM.MARIST DOT EDU
        Subject:        Reclamation problem - tape robot gets full!

        >Hello,
        >
        >we need help about reclamation in our tape pool (IBM 3494).
        >
        >We run ADSM 3.1 on MVS server.
        >
        >It seems that the reclamation processes do not run logically
(we applied the
        >patch concerning aggregates):
        >
        >as you can see below, adsm reclaims tape 3180 by writing on
tape 3276, which
        >becomes full, then asks for tape 3323 (scratch tape).
        >
        >  Tape 3323 is almost immediately deleted!! and tape 3510 gets
asked
        >
        >Next reclamation process: adsm reclaims tape 3276 which was
just written to!!
        >
        >This kind of loop continues then. We have recl=67%.
        >
        >Does anybody experience the same situation?
        >
        >IBM, is there a bug, will it be corrected in the awaited
definitive patch
        >concerning the existing error in reclamation (logical spaces)?
        >
        >Please help, our tape robot is going to be full in a few days!!
        >
        >

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