Thanks for such a quick response Stefan, it is greatly appreciated!

>>>>> On Fri, 23 Jan 2004, "Stefan" == Stefan G. Weichinger wrote:

  Stefan> Have a look in the amanda-manpage. In the section DISKLIST
  Stefan> FILE you will find the explanation (yes, they are kind of
  Stefan> labels).

I actually found this yesterday, but mis-read it.  I re-read it again 
just after I sent my earlier mail and read more carefully.  I think I 
now understand.

  Stefan> Something like:

   jpt /u1/backup/all /u1/backup {
     # Exclude everything that's explicitly named
     # if we add a directory later, but forget to add it here
     # it will be picked up automagically
    non-user-high-tar
   exclude list "/usr/local/etc/amanda/daily/backup-excludes"
  } 1

I had thought of this, but then realized that this would also require
the maintenance of these files on each client (not a big deal, as it's
all in cvs, I just need to restructure some things).

  Stefan> And after that you list the dirs to be INcluded like this:

So you EXclude before you INclude? Or does it not matter?

  Stefan> The dumpsize for /u1/backup should be nearly nothing and so
  Stefan> you can check if your excludes work out fine. If it gets
  Stefan> bigger maybe some new directory has been created

This is exactly the behavior I am trying to achieve.

  Stefan> or some exclude does not work as intended.

This I'd like to avoid ;)

So, assuming for the moment that there is no practical difference 
between using "exclude list <file.name>" and
"exclude [append] pattern" I'll stay with the latter for now for 
testing purposes and clarity, and move to the former at a later time
for ease of maintenance.

To clarify, do I want:

 jpt /u1/backup/all /u1/backup {

  exlude <all subdirs which follow as separate "disks" >

 } 1

Followed by:

 jpt /u1/backup/<subdir> /u1/backup {
  include <specific subdir>
 } 1

or:

 jpt /u1/backup/<subdir> /u1/backup {
  include <specific subdir>
  exlude <everything else>
 } 1

?

Thanks again!
-- 
Seeya,
Paul

GPG Key fingerprint = 1660 FECC 5D21 D286 F853  E808 BB07 9239 53F1 28EE

         If you're not having fun, you're not doing it right!