In <199803061903.OAA223912 AT acs3.bu DOT edu>, on 03/06/98
   at 02:03 PM, Richard Sims <rbs AT BU DOT EDU> said:

>>Also, let's not forget that remote host disks do go bad, and that
>>performing a full restore from an "incremental forever" backup set
>>will usually result in a extraordinary number of tape mounts and
>>a correspondingly long restore window. Periodic full backups help
>>alleviate this problem.

>Your criticism is ignoring all the ADSM architectural features which deal
>with this, such as Collocation By Filespace and Tape Reclamation. If you
>want to base recovery on traditional full + incremental backups, you should
>not bother with ADSM, as it's just a waste of money for the more basic
>things you want to do.  You may want to review your enterprise's needs, as
>well as the full repertoire of ADSM's capabilities, and decide direction
>based upon that larger view.  Also review the ADSM-L archives, as this
>issue has been extensively discussed in the past.
>     Richard Sims, BU

I agree with Richard - very well said.

I would like to add that ADSM is a very big thing and it takes  a while to
get your arms around it.  In fact it isn't very hard at  all to do periodic
full backups, and you don't need to play with the  domains or management
classes at all.  So if you have the media and database space here is how I
think it can be done (disclaimer -
I don't do this, I am comfortable with the incremental forever model).

Just rename the filespaces with the server 'rename filespace' command. The
next backup will be a full backup.  The renamed filespace will be a frozen
snapshot, the active set will never change, it will last until you delete
the filespace.  I am sure scripts could be written and it shouldn't be hard
to implement a scheme like this.

If you turn on collocation by filespace, the new full backup will be stored
on different tape media.
--
-----------------------------------------------------------
-----------------------------------------------------------
Bill Colwell
Bill Colwell
C. S. Draper Lab
Cambridge, Ma
bcolwell AT draper DOT com
-----------------------------------------------------------
=========================================================================