[ejabberd] Strange crash on ejabberd slaves

Christopher Zorn christopher.zorn at gmail.com
Mon Jun 22 23:44:24 MSD 2009


I am running a cluster of ejabberd servers. Normally, when a crash happens,
I see an erl_crash.dump or mnesia has a core dump. These are fine because
there is a way to figure out the problem.

But I recently just had a crash where the ejabberd process was still running
but I could not connect to port 5222. I also could not connect with
ejabberdctl/remote shell. There was nothing in the sasl log or ejabberd log
to indicate the reason for failure. There was no crash dump or core dump. I
then tried to restart and the sasl log indicated that everything started up
correctly. The listenters were running but I could not connect to port 5222.
It was not until I restarted the master node that I was able to get the
slaves back up.

I am currently trying to reproduce this bug so I can get some better
information about it. I just wanted to send this email to see if anyone had
an idea what was going on here. Maybe something I should check that I may
have missed?


I am running erlang R.13b and ejabberd svn revision 2118

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.jabber.ru/pipermail/ejabberd/attachments/20090622/0419d4e5/attachment.html>


More information about the ejabberd mailing list