Hi,
I've been using ldap for a few years now, but I've come to the realization that I have a lot of mysql databases lying around that should be integrated and unified. Its time to move to something more robust.
I'm using openldap-24 and am thrilled with the responsiveness and extensibility of the software, but I'm curious about using mysql as a back end instead of bdb.
I've been able get unixODBC up and running without any problems, but I'm stumbling a bit over the 10,000 foot view on metadata and how to generate metadata to insert a samba 3 object.
After skimming for a few days, I found this lovely overview (http://www.openldap.org/faq/data/cache/978.html) which didn't inspire a lot of confidence in how easy it would be to support abstract schemas that we might want to host on this system.
From this discussion and subsequent threads (http://www.openldap.org/lists/openldap-software/200511/msg00504.html) , I'm concerns about the practicality and that I might have unrealistic expectations for being able to deploy 250k records on my mysql cluster to service a series of ldap servers over back-sql.
Does anyone run back-sql in production, and what kind of records on what scale of system are you using ?
I have high confidence that I want to use ldap, but I'd rather right provisioning software that talks to a mysql database than to talk to the ldap database. To make things more complicated, I suspect we will be moving to oracle in the next few years.
Does anyone have any suggestions on a 10000 foot view of metadata ?
Thanks.