We have two LDAP servers setup, a master and a replica. From what I understand we should be able to take down the replica for maintenance and the servers should start talking to the master as a backup.

Unfortunately it seems that even though the replica server is off the servers still try to talk to it and wait 30 seconds before giving up. I've found the "ldap_connect_timeout" and "ldap_read_timeout" localconfig options and set them to 5000 instead of 30000 to see if this helps.

Is there something we should do to stop the servers from trying to reach the replica while it's down for service? (Other than edit the localconfig and restart the services)