After discussing the problems with others, it is determined
that node 1 is a problem.

I was able to work around this problem by naming my nodes 10, 20, 30.
After a failover I was able to add Node10 back using the same
procedure that failed when I used Node 1.

Also, check your slonik script and configuration.  If node 2
is your origin node and your configuration says it is node 1,
the error message makes perfect sense.

--elein 

On Tue, Aug 30, 2005 at 03:35:15PM +0200, St?phane P?trot wrote:
> Hello,
> 
> I am doing some tests about failover process for a simple cluster :
> node1 = master
> node2 = slave
> 
> I run the perl script  : slonik_failover node1 node2 where node1 is the dead
> one and node2 is the backup.
> Everything works fine.
> 
> Then, I want to restore the previous configuration and I don't know exactly how
> to do. I have read that one solution is to restore the database of node1 from
> scratch as a subscriber. But when I run the script (on node2) :
> slonik_subscribe_set set1 node1
> I have an error telling that it is impossible to subscribe to the set with the
> same origin and subscriber (node1 in my case), because node1 is the master in
> the configuration file that I use. Do I have to use another configuration file
> after running the failover script, telling that the new master is node2? And
> then doing switchover from node2 to node1?
> 
> Thanks
> St?phane

> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general