Re: question on where to expire entries in server code

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

At 10:02 PM 2002-10-09, Gerald Carter wrote:
>Greetings,
>
>I'm working though the slapd source (2.1.5 for the moment) and
>am trying to decide the best place to expire directory entries
>based on the modifyTimeStamp attribute (after so many seconds or 
>what not).  
>
>The reason is that i'm using openldap to hold publishing information 
>for current devices installed on the network.  A device is responsible
>for periodically refreshing its entry somehow.

You might look at RFC 2589.

>Initially i thought that doing this in the bdb backend would be best,
>but then considered that if it was implemented as an ldap control
>on a search request, then it should be done above the backend 
>interface.  Any thoughts on this?  Am I moving in what would seem to a 
>sane direction?  Has this already been done some other way that i've
>missed? 

You might look at the FTree backend (for 2.0).
  http://www.openldap.org/its/?findid=1499

IIRC, FTree doesn't provide persistence for dynamic objects.
But, depending on your application, that may not be needed.

Kurt