ADSM-L

Re: How would ADSM deal with a large numbered workstation deployment?

1996-12-11 12:06:21
Subject: Re: How would ADSM deal with a large numbered workstation deployment?
From: "Prather, Wanda" <PrathW1 AT CENTRAL.SSD.JHUAPL DOT EDU>
Date: Wed, 11 Dec 1996 12:06:21 -0500
We are using ADSM to backup network servers, and also in the planning
stage for backing up desktops, up to 500.

We also centrally manage Windows software installs so that most of our
users are at the same level of software for Windows and a suite of
office applications.

In our shop, your estimate of 6000 files per desktop is best-case for
Win95, and too low by at least  2000 for WinNT.  (I expect it to get
significantly worse when Office 97 becomes available..)
Best I can calculate, my ADSM DB will be up to 6 GB by the end of the
year, for about 350 clients.

What I am contemplating at the moment is putting the dsm.opt for all the
clients on a network server where we can force-update their exclude
lists for them.  Or maybe having only 5 or 6 dsm.opt files and having
all clients use one of them so we only have to maintain 5 or 6 exclude
lists.

And it's very tricky to build those exclude lists, because you can't
usually exclude by directory -- Windows puts user customization files
everywhere.

It would help tremendously if we had a "global" exclude list capability
where the administrator could build a central exclude list (perhaps for
each management class) that would be used in addition to the client's
exclude list.


>----------
>From:  Dennis Taylor[SMTP:TAYLORDC AT gunet.georgetown DOT edu]
>Sent:  Wednesday, December 11, 1996 9:01 AM
>To:    Multiple recipients of list ADSM-L
>Subject:       How would ADSM deal with a large numbered workstation 
>deployment?
>
>We have been using ADSM for more than 3 years to backup a small (less
>than 50) collection of Unix, NT, Windows, Novell--servers &
>workstations.
>
>My organization is planning a rapid deployment of a large number of
>Windows (TBD 95 or NT) based workstations--500 in the first wave alone.
>The workstations will be identical, all loaded with the same template of
>office suite and other applications.
>
>My question is how (if at all) to propose a backup solution using ADSM.
>
>The problem, which has been discussed on this list previously, is the
>large amount of redundant data.
>
>500 workstations at say 6000 files each and 600 MB each will require on
>the order of 2GB ADSM database capacity and nearly 300 GB storage pool
>capacity.
>
>All of this for non-original/non-unique data.  The suggestion on this
>list in the past was to just ramp up HW and deal with it...  I just
>don't think it would fit in this case.  We would tie up a relatively
>large automated tape library for nothing.
>
>I would guess that I could exclude the template files (the 6000).  But
>if the user downloads his/her own patches to the OS or application code
>we wouldn't pick it up.  Perhaps dynamically build a specific exclude
>list (but the I am not sure scheduled backups would reread this)?  Use
>the API (no)?
>
>Comments or suggestions would be appreciated.
>
>Dennis Taylor
>taylordc AT gunet.georgetown DOT edu
>
<Prev in Thread] Current Thread [Next in Thread>