-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I tried to add a set with some tables and sequences to an already existing
replication.

Trouble is I messed up numbering my sequences, and then I obtained some error :

ERROR  remoteWorkerThread_1:
"select "_slonturf".setAddSequence_int(3, 39,
"public"."rf_ihm_genset_id_ihm_genset_seq"', 'sequence
public.rf_ihm_genset_id_ihm_genset_seq')"

PGRES_FATAL_ERROR ERREUR: Slony-I: setAddSequence_int(): sequence ID 39 has
already been assigned
remoteWorkerThread_1: data copy for set 3 failed - sleep 60 seconds


After I understood why it broke the replication, I dropped the new set from
replication.
And I recreated it, with correct numbering of objects in it.

2 of my nodes didn't resynchronize.

I tried to drop the set again, but now I do see that error in logs :

slon[18542]: [72-1] 2008-12-10 13:29:29 CET ERROR  remoteWorkerThread_1: node
- -1 not found in runtime configuration
slon[18542]: [73-1] 2008-12-10 13:29:29 CET WARN   remoteWorkerThread_1: data
copy for set 3 failed - sleep 15 seconds

On my master, set 3 and related objects don't exist anymore.
On the two desynchronized sets, set 3 still exists, but sl_table and
sl_sequence don't know about objects in the set.

For information, I'm on PG8.2.11, with Slony 1.2.15.
Replication goes this way : 1->11->71->72. 71 and 72 can't synchronize anymore.

On node 71/72, the set still exists in sl_set and sl_subscribe. Is there a way
I could purge nodes from any reference to that set ?

Thanks by advance.

- --
Stéphane Schildknecht
PostgreSQLFr - http://www.postgresql.fr
Dalibo - http://www.dalibo.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFJP7wLA+REPKWGI0ERAiKbAKDDJb+oLTBi0787aRuBE7e7EN4W3ACeO9q+
+vGQQ8ndG6cA7tcJS7zdbRQ=
=SRZu
-----END PGP SIGNATURE-----