We want to use AD as a regular LDAP for storing and accessing our application data in fast and reliable way. We have two physical server and want to organize some sort of clustering for AD and its data. What is the best-practice for such task?
-
1The best practice for "clustering" AD is, I'd say, "Don't". Just use all the machines you would have had as cluster members as separate domain controllers.– Rob MoirJun 17, 2011 at 21:51
-
Is your approach correct when I use AD just as regular LDAP? (like in serverfault.com/questions/195025/…)– SashaJun 17, 2011 at 21:54
-
2I think so. You cluster to improve availability right? AD already has a mechanism for that. Don't think of the availability of this or that server, think of the availability of the service.– Rob MoirJun 17, 2011 at 23:30
1 Answer
AD is already clustered. You bind to the domain using the correct entries that AD has stored in DNS. Edit - to clarify :
You don't do anything to cluster AD. The process of dcpromo to make a server a DC will automatically make the new DC take part in AD replication. That keeps the AD database synced. You don't need to do anything to make you application failover-aware either; you bind to the domain, not a single serve
Having said that, it's not a great idea to store much of your own data in AD's schema. You may want to make your own partition, or use LDS (formerly ADAM.)
-
Thank for answering. But what does it mean "already clustered"? I have two servers, what do I need to make a AD clustered?– SashaJun 17, 2011 at 21:50
-
3You don't do anything to cluster AD. The process of dcpromo to make a server a DC will automatically make the new DC take part in AD replication. That keeps the AD database synced. You don't need to do anything to make you application failover-aware either; you bind to the domain, not a single server.– mfinniJun 17, 2011 at 21:54