[ejabberd] ejabberd servers recovery from net partition

Zbyszek Żółkiewski zbyszek at toliman.pl
Tue Mar 13 01:32:40 MSK 2007


well thats one of reported issues. I think this is erlang issue, but
probably it can be workaround by "building-in" some kind of heart-beats.
Probably Aleksey or Mickaël Rémond can write some more about that problem.

btw. when 2 or more nodes have network issues between them - it can lead to
very serious problems, including database inconsistencies (mnesia), presence
problems etc...

On 3/12/07, Igor Goryachev <igor at goryachev.org> wrote:
>
> "Zbyszek Żółkiewski" <zbyszek at toliman.pl> writes:
>
> >> I'm running ejabberd in a clustered configuration on two seperate
> hosts.
> >> Over the weekend both servers stopped acknowledging each other (net
> >> partition). What is the standard recovery procedure for such an
> instance.
> >>
> >> Note: the two boxes can now ping each other
> >
> > cluster restart. it is known issue, when link between min. two ejd nodes
> is
> > lost - nodes cannot reconect - search this group for details.
>
> The same failure sometimes occurs with my servers. Is this an ejabberd
> or erlang issue itself? Is there a solution instead of cluster restart,
> which seems to be redundant?
>
> I have found
> http://lists.jabber.ru/pipermail/ejabberd/2006-September/002145.html,
> but there is no concrete answer or solution/workaround.
>
>
> --
>     Igor Goryachev              E-Mail/Jabber: igor at goryachev.org
> _______________________________________________
> ejabberd mailing list
> ejabberd at jabber.ru
> http://lists.jabber.ru/mailman/listinfo/ejabberd
>



-- 
pozdrawiam,
Zbyszek Żółkiewski
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jabber.ru/pipermail/ejabberd/attachments/20070312/5b9f60c3/attachment.htm 


More information about the ejabberd mailing list