Clustering with LDAP doesn't allow login to second node if first node is down

I have succesfully cluster several nodes together and they support LDAP authentication. Problem is if the first node is down and the second node restarts while the first node is still down, I cannot log into the second node using a jabber client. Further to that, it appears that the second node is not listening on any ports. It's like it hasn't loaded the config. Is this normal behaviour or have I configured something wrong? The second node starts ejabbered with the "-mnesia extra_db_nodes "['ejabberd@first']" line added in. I'm guessing that because the first node is down, the second one is unable to load the config properly because it can't add the first node. Any help is appreciated.