we seem to be having trouble getting s2s to work properly in our organisation. We have set up 2 Wildfire servers on our internal network and named them (i.e. set xmpp.domain to) the respective host name, e.g. server1 & server2. We have integrated them with our AD and client connectivity (c2s) is working fine. Now, if I log in to server1 with Psi I seem to be able to browse and see the services of server2 listed but I can’‘t actually use them. User search for example is not possible as the option on Psi is greyed out. I’'m wondering if this has something to do with a name resolution/dialback problem. We have not set up any SRV records for our servers, nor are we specifying the FQDN in the server setup (i.e. server1.internal.net, server2.internal.net). Is this required? Below is the output on the XML console in Psi for when I try and use the search service on server2 (bddh01-s0001).
Any ideas/help on this would be greatly appreciated.
I’'m wondering if this has something to do with a name resolution/dialback problem.
Looks like you’‘re correct in that sense. In my case, I fix the problem by specifying all sub-domains/component names in Wildfire to point to the same IP of the main domain ( xmpp.domain ). To Wildfire, I believe somehost and search.somehost don’'t necessary interpret to the same host because you can have 2 completely different Wildfire servers each with somehost and search.somehost names. However, when you define both names to refer to the same Wildfire, s2s will maintain a single connections between servers that carry both names. You could verify this with “netstat -tn | grep 5262”.
We have not set up any SRV records for our servers, nor are we specifying the FQDN in the server setup (i.e. server1.internal.net, server2.internal.net). Is this required?
It’'s not required but there has to be some way to resolve the names. At least, an /etc/hosts entry would do.