I'm trying to replicate two databases using (partially) the following to
set up slony:

...
# Set 1, for schema alpha
create set (id=1, origin=1, comment='Alpha');
set add table (set id=1, origin=1, id=1, fully qualified name =
'alpha.category', comment='');
....
# Set 2, for schema beta
create set (id=1, origin=1, comment='Beta');
set add table (set id=1, origin=1, id=1, fully qualified name =
'beta.category', comment='');
....
# Set 3, for schema gamma
create set (id=1, origin=1, comment='Gamma');
set add table (set id=1, origin=1, id=1, fully qualified name =
'gamma.category', comment='');
....

Running the script i get this:

<stdin>:48: PGRES_FATAL_ERROR select "_slonyTest".setAddTable(2, 1,
'beta.category', 'category_pkey', '');  - ERROR:  duplicate key violates
unique constraint "sl_table-pkey"
CONTEXT:  SQL statement "insert into "_slonyTest".sl_table (tab_id,
tab_reloid, tab_relname, tab_nspname, tab_set, tab_idxname, tab_altered,
tab_comment) values ( $1 ,  $2 ,  $3 ,  $4 ,  $5 ,  $6 , false,  $7 )"
PL/pgSQL function "setaddtable_int" line 74 at SQL statement
SQL statement "SELECT  "_slonyTest".setAddTable_int( $1 ,  $2 ,  $3 ,
$4 ,  $5 )"
PL/pgSQL function "setaddtable" line 37 at perform

The schemas alpha, beta, gamma share the same table names and index
(PK,FK) names too. So fully qualified names of PK indexes are like this:

alpha.category_pkey
beta.category_pkey
gamma.category_pkey

So, i'm guessing that slony has a UNIQUE constraint on indexes names,
and they are not stored with their fully qualified names (that is,
including the schema they belong to) within the sl_table-pkey. I'm a
right?

Carlos Gaona U.
Confianze