> object from all TSM operations, the product unfortunately departs from
> obviousness here, requiring the .archive
> qualifier to exclude from archiving as well, as has been consistently
> explained in the client manuals.

While the manuals do state that Exclude will exclude files from backup, and
independently the manuals do state that Exclude.archive will exclude files from
archive, it does not make it clear that Exclude by itself will *not* exclude
files from archive, and the fact that the Exclude.backup function exists, to my
mind, implys that Exclude (by itself) would not limit itself to backups, or in
other words it would successfully exclude archives.

For quite some times I've made an assumption based on this which turned out to
be a poor assumption.  This was only clarified by the new 5.3 install which
separates exclude.backups from exclude.archives in it's default option file
creation.

I agree whole heartedly that this is probably a side effect of 'evolving
features'.   Kind of like Policy Domains, where the single biggest use of
Policy Domains these days is to separate Agent data from regular backup data,
as opposed to what a Domain was originally designed for...

regards,

Paul


Quoting Richard Sims <rbs AT BU DOT EDU>:

> On Feb 11, 2005, at 7:27 AM, Paul Fielding wrote:
>
> > ...My belief has always been that exclude would exclude both backups
> > and archives,
> > which would negate the need to exclude the same filespec with both
> > exclude.backup and exclude.archive. ...
>
> While an unqualified Exclude should intuitively exclude a file system
> object from all TSM operations, the product unfortunately departs from
> obviousness here, requiring the .archive
> qualifier to exclude from archiving as well, as has been consistently
> explained in the client manuals. Products sometimes have warts like
> this due to the way they evolve from initial design objectives, as in
> Backup being the initial role of the primordial product.
>
>     Richard Sims
>


----------
paul AT fielding DOT ca
http://www.fielding.ca

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/