With all due respect, I supported an international account exactly this way; as 
an interim solution (while they ordered up a second tape drive) at 3 of their 
12 sites, using policyset swap to re-drive the full-backups (ie, change 
copygroup to "absolute", from "modified") was just what we needed, so they 
could run (almost) lights-out for a month at a time, using incrementals after 
each monthly full.

It's not the best solution, but with only a single drive, this gets 3 copies of 
data on tape (for files which exist between any 2 full-backups).  It clearly 
defeats the purpose of TSM's design, but when a customer "screws up" and fails 
to get proper equipment on first attempt, it saves the TSM decision, while 
scrambling to get the 2nd drive for true, lights-out/normal operation.

If you've got the bandwidth to run the full-backup over a weekend, this fits 
the bill;  if not (due to multiple nodes), then you'll need to work harder and 
support multiple node-names per client machine... one for daily INCR, one for 
weekly INCR --- Hmmm, maybe that's a better scenario for your situation, eh?  
(My customer's scenario worked best because it allowed us to do en-mass swap of 
scratch tapes for old backup tapes... which facilitated remote restore from the 
help desk, transparently to the help-desk folks!) 

HTH!


Don France
Technical Architect -- Tivoli Certified Consultant
Tivoli Storage Manager, WinNT/2K, AIX/Unix, OS/390
San Jose, Ca
(408) 257-3037
mailto:don_france AT ayett DOT net (change aye to a for replies)

Professional Association of Contract Employees 
(P.A.C.E. -- www.pacepros.com)



-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]On Behalf Of
Steven Pemberton
Sent: Thursday, May 22, 2003 5:09 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Question on defeating TSM strengths: due to budget
constraints << solved


On Friday 23 May 2003 07:41, William F. Colwell wrote:

No, no, no! This is _so_ wrong I don't know where to begin... :)

> I think this is one of the rare cases where policysets are useful.
> Policysets can implement time based policies.
> Just copy your STANDARD policyset to WEEKLY.  Edit the
> STANDARD mgmtclas (copygroup) to have the new parameters, i.e.
> mode=absolute.

Although multiple policy sets are *sometimes* useful, this isn't one of those 
occasions.

Muliple policy sets are only useful when you are using the configuration 
manager to replicate a policy domain to another TSM server. You could then 
define two different policy sets on the configuration manager, replicate the 
policy domain, and then activate a different policy set on each TSM server.

> Now setup  an admin schedule to activate the WEEKLY policyset at the right
> time, and another to activate STANDARD.  This works because the client
> queries the server for policy information at the start of each backup,
> which is different for the problems the scheduler has with include/exclude
> lists which need a scheduler restart to take effect.

If you activate a different policy set, then the management class/copy group 
parameteres in *THAT* policy set will control how the data is retained. Only 
the policy set which is active at the time *EXPIRE INVENTORY* is run will 
control whether the data is retained or expired. It doesn't matter (much) 
which policy set was active at the time of the client backup.

> >Ken_Sedlacek AT KYRUS DOT COM said...

> >I was saying my rosary this morning, and by divine revelation (so to
> >speak), just realized that all I need to do is make a schedule with the
> >new mgmt class. This is so simple, I could just spit!

Remember, if you change the management class being used for backup objects it 
will rebind all the existing backup objects to the new management class.

> >1) How do I get the client servers to execute this FULL_Backup Mgmt class
> >on a weekend schedule, while maintaining the STANDARD Mgmt class for M-F?

There are several ways to do something like this. The simplest is probably to 
just define two client accounts per node. Create two client scedules 
(possibly in different policy domains...) and backup the node (using the 
"normal" management class) every day, and backup the "other" node (using the 
"special" management class once per week (or so).

You *really* need to buy a second tape drive. :)

PS/ You *REALLY* need to buy a second tape drive. :)

Regards,
Steven P.

-- 
Steven Pemberton                                      Mobile: +61 4 1833 5136
Innovative Business Knowledge                   Office: +61 3 9820 5811
Senior Enterprise Management Consultant    Fax: +61 3 9820 9907