[ejabberd] Error upon login after primary node stopped.

tty at SAFe-mail.net tty at SAFe-mail.net
Tue Mar 8 00:23:27 MSK 2005


I have a Jabber domain with two ejabberd nodes on two different machines. The secondary node replicates all the table entries of the primary (using mnesia:add_table_copy). I can login to the secondary node, talk to users on the primary, and interact via the MUC.

However, if I shut down the primary node (init:stop) and restart it my client is unable to browse services on the primary. 

If I then logoff and login I get the following error:

     +++++++++++++++++++++++++++++++++++
=ERROR REPORT==== 7-Mar-2005::14:28:52 ===
Discarding message (2,3){route,{jid,"totem","
jabber.kataragama.com","Psi","totem",
"jabber.kataragama.com","Psi"},
{jid,"totem","jabber.kataragama.com",
"Psi","totem","jabber.kataragama.com",
"Psi"},{xmlelement,"iq",[{"id","aab1a"},
{"type","result"}],[{xmlelement,"query",
[{"xmlns","jabber:iq:roster"}],[]}]}}
to a process in an old incarnation of this node

     +++++++++++++++++++++++++++++++++++

Browsing to the admin console shows I am login but my  client hangs continously showing a login attempt. I tried with psi 0.9.2, ejabberd 0.7.5 and otp_src_R10B-3. 

Using tkabber 0.9.7 it indicates that I'm not login and the  console of the secondary node shows an attempt to login to the primary. i.e

   totem at jabber.kataragama.com 

instead of

   totem at jabber.noddle.com

even though the configuration of tkabber is to use jabber.noddle.com.

Do I have something mis-configured ? I am looking at having 2+ nodes in the Jabber domain which can be added/remove on the fly.

Thanks

totem





More information about the ejabberd mailing list