But it seems to make much sense: In my experience if you configure multiple LDAP servers, the NSS resolver always uses the first configured server as long as it's reachable; even if it's not, the first configured server is tried first. After connection timeout the second server is tried... We had dhad a case when all LDAP operations were heavily delayed when two out of three servers had failed (due to rebooting the machine).
That still has absolutely zero to do with MMR replication. There are useful and valid reasons for using load balancers across a replicated pool. Your reponse however indicates a lack of comprehension of what Mirror Mode entails.
--Quanah -- Quanah Gibson-Mount Product Architect Symas Corporation Packaged, certified, and supported LDAP solutions powered by OpenLDAP: <http://www.symas.com>