On Wednesday 03 September 2008 12:11:44 pm David Lee Lambert wrote:
>
> X, Y and P have exactly the same schema already.
>
> X has stale test data.  Y has an up-to-date exact replica of the stale test
> data.
>
> P has production data.  I want Y to have an up-to-date exact replica of the
> production data.

I was thinking through this problem some more,  and it seems like the best 
solution is to have two different sets, one for each potential master;  that 
way I can switch the slave back and forth between "production" and "test" 
data with just a "slonik_unsubscribe_set" followed by 
a "slonik_subscribe_set".  However,  when I try creating the second set ofer 
the same set of tables but with a different origin,  I get an error:

<stdin>:19: PGRES_FATAL_ERROR select "_replication".setAddTable(7, 
101, 'public.addresses', 'addresses_pkey', 'Table public.addresses with 
primary key'); - ERROR:  duplicate key violates unique 
constraint "sl_table_tab_reloid_key"

Is that constraint really necessary for Slony-I to function correctly?  
Shouldn't it be possible for a table to be in more than one set,  as long as 
a client is never subscribed to two overlapping sets?



-- 

David L. Lambert
  Software Developer,  Precision Motor Transport Group, LLC
  Work phone 517-349-3011 x215
  Cell phone 586-873-8813