subtree repl (was: should one single search be able to encompass several databases?)

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 07:59 AM 7/19/2001, Pierangelo Masarati wrote:
>Stig Venaas wrote:
>
>> I'm not sure if the subject is clear. The situation I'm thinking of is
>> something like this:
>>
>> You create one database with suffix say dc=test,dc=com, and another with
>> suffix dc=hr,dc=test,dc=com. If someone does a subtree search with base
>> dc=test,dc=com, should the search also be done in the dc=hr,dc=test,dc=com
>> database? I think maybe it should, but it doesn't now, and I suspect it's
>> hard to implement. Should we implement this, or should we simply say that
>> people must put it all in one database?
>
>BTW, the I guess the most important reason for people to use different
>databases for subtrees of the same naming context is partial replication.
>In this sense, it might be easier to implement "selective" replication based
>on the naming context: if we add a field "suffix=<naming context>" to
>the "replica" statement, then only the entries whose dn has
><naming context> as ancestor could be easily replicated.
>
>Comments?

In addition, I'd like to see the entries written to the replog filtered by
some means...  use of ACLs to control partial/sparse replication is one
approach which would be interesting.

However, I have little desire to put to much investment of time into
replog/slurpd as I rather have an LCUP based solution...

Kurt