[ejabberd] Re: strange dependency

Magnus Henoch mange at freemail.hu
Tue May 31 11:19:43 MSD 2005


Felipe Brito Vasconcellos <fbv at trf4.gov.br> writes:

> Hello.
>
> I have done the clustering steps and got success.
>
> But, now the second node seems to be dependent of the first one.
>
> I mean, if I stop (with my kill method...... never got problem with it 
> in two or three months of use, but I will change this as soon as 
> possible) both nodes, and then try to start just the second (new) one, 
> the service just don't work right. I mean, I  *need* to start the first 
> AND the second node to get the service up. Is this correct? I think 
> there must be a way to catch out the first node from this cluster, 
> because all the tables were copy to both nodes.

It seems that you have to use mnesia:force_load_table [1] for this
situation, as the nodes start up with caution - to avoid ending up
with divergent databases, they want manual confirmation that they may
go ahead without the other nodes.

Magnus

[1] http://www.erlang.se/doc/doc-5.4.3/lib/mnesia-4.2/doc/html/mnesia.html#force_load_table/1



More information about the ejabberd mailing list