ejabberd on several hosts interconnected with several virtual hosts

ok, guys,

first, great work about ejabberd, i like it.

now, i would like to accomplish following tasks:

(1) I have several erlang hosts, i would like to get them working together.
(2) On this hosts, i serve first of all one domain.
(3) Now, i want to add on the one of them the second vhost
(4) i would like to have some of them serving domain 1, some other of them domain 2. some of them both.

I got the first and second task working. Now i would like to manage third and fourth task, but i get stuck. Even if i just try one box serving two domains, i get stuck, because on one vhost i can register, on the second i am not allowed to.

i have the two domains on the masterbox in the hosts-directive, and for each domain a config_host-directive. If i register the first user on one of the domains, this one will grant me further user registration, the other will deny.

any hints?

Sincerly

DevNiv

progress...

ok, i am a little further... i experience problems with some transports... always after a while.

PyICQt-Contacts claim to be PyMSN-Contacts... PyAIM fights against PyICQ... thats strange....

another progress

nice... the transports shuffle spool-files... thats great, i have msn transport-spool-configs in my icq transport spool directory...

did anyone have these experience, too?

do I have to ask the py$transport guys to help me? or is that an ejabberd-thing?

Re: another progress

DevNiv wrote:

nice... the transports shuffle spool-files... thats great, i have msn transport-spool-configs in my icq transport spool directory...

Each transport is a different program that connects to ejabberd using the same or a different port, right? How can ejabberd modify spool dirs of programs?

Try putting each transport on a different port, and with a different password. It may be that ejabberd was sending messages to the wrong transport. However, considering this is the first time I read about such I bug, it may be a misconfiguration.

Re: another progress

badlop wrote:

Each transport is a different program that connects to ejabberd using the same or a different port, right? How can ejabberd modify spool dirs of programs?

they all use the same ports and passwords... for that event last weekend i disabled all but ICQ... with that ejabberd was stable enough.

badlop wrote:

Try putting each transport on a different port, and with a different password. It may be that ejabberd was sending messages to the wrong transport. However, considering this is the first time I read about such I bug, it may be a misconfiguration.

well i just checked it out at the moment of realizing that. I even put all the transports into completely other directories... but they are connecting with the same password and the same port...

I send you my ejabberd.cfg, if you want to analyze it. Expect an answer for the next weekend. I am busy right now.

Oh, another thing... i got an icq contact claiming to be a jabber-service (just like pyicq does) the last days i re-initialized a fresh spoolfile. It vanished after i removed it from my roster and it hasn't apeared again,yet... Stay tuned *G*

.

.

Re: another progress

DevNiv@www.ejabberd.im wrote:

well i just checked it out at the moment of realizing that. I even put all the transports into completely other directories... but they are connecting with the same password and the same port...

You need at least a different port for each transport, and this needs to match the ejabberd configuration file. Check ejabberd's documentation for an example confgiration.

ok, that different port did

ok, that different port did it... it runs now stable, not shuffeling the spoolfile-content around... thx for that hint :)

Syndicate content