ADSM-L

Re: Maxscratch in a 3494

1999-07-15 16:51:35
Subject: Re: Maxscratch in a 3494
From: Dwight Cook <decook AT AMOCO DOT COM>
Date: Thu, 15 Jul 1999 15:51:35 -0500
     Looks like you might have "collocation yes" or even more so
     "collocation file"  ADSM's selection process will be to first look for
     an available "filling" tape that meets the needs (based on
     collocation) if it doesn't find one, then it grabs a scratch (if a
     scratch isn't available and based on collocation being something other
     that no, it will grab a filling tape)

     Are you going straight to tape ? ? ?

     I like to front end buffer my inbound traffic with a diskpool... I
     make the diskpool large enough to handle a single day's inbound
     traffic... this also lets me take my concurrent inbound client
     sessions to about 15+ times the number of tape drives I have on the
     adsm server...  (and dasd is cheap anymore)  Oh, and one server had
     its clients pull a fast one on me and where I only have a 300 GB
     diskpool, lately they've been throwing an average of about 550 GB a
     night at me... (the diskpool starts automigration to tape at 80% and
     never hits 100% due to the bleed-off)

     Dwight




______________________________ Reply Separator _________________________________
Subject: Maxscratch in a 3494
Author:  Jones4A (Jones4A AT KOCHIND DOT COM) at unix,mime
Date:    7/15/99 3:30 PM


Any one have any good tips on controlling the amount of tapes TSM (a.k.a.
ADSM) grabs when it needs a tape?

The manuals recommend setting maxscratch equal to the total number of cells
available, but that leads to a lot of 0.? filling volumes.

I have tried reducing this number, only to have some of the backups fail
with "no more space left in storage pool" messages.  NOT good, when their
really is space available.

Any suggestions appreciated.

Anthony A. Jones
316-828-2432
Business Continuity & Data Storage Management
Koch Industries, Inc.
<Prev in Thread] Current Thread [Next in Thread>