I have a bit uncommon configuration now.
Master: postgresql 8.3
Slave:   postgresql 8.4
slony:   version 1.2.17

Any execute script surely break replication set. Even if i submit
something like 'select 1' via execte script replication will be broken
down unrepairable. Only way fix it is drop/create/subscribe slave node.
restarting slon doesn't help at all...

Here is slony log from slave database:

2009-12-08 09:11:43 MSK INFO   prepare for DDL script - set:1 onlyonnode:-1
2009-12-08 09:11:43 MSK ERROR  remoteWorkerThread_1: "select
"_slony".ddlScript_prepare_int(1, -1); " PGRES_FATAL_ERROR ERROR:
Slony-I: alterTableRestore(): Table "public"."access_type" is not in
altered state
CONTEXT:  SQL statement "SELECT  "_slony".alterTableRestore( $1 )"
PL/pgSQL function "ddlscript_prepare_int" line 46 at PERFORM
2009-12-08 09:11:43 MSK ERROR  remoteWorkerThread_1: DDL preparation
failed - set 1 - only on node -1
2009-12-08 09:11:43 MSK DEBUG1 slon: retry requested
2009-12-08 09:11:43 MSK INFO   remoteListenThread_1: disconnecting
from 'dbname=plus1 host=192.168.1.137 port=5432 user=slony'
2009-12-08 09:11:43 MSK DEBUG1 syncThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 localListenThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 remoteListenThread_1: thread done
2009-12-08 09:11:43 MSK DEBUG1 main: scheduler mainloop returned
2009-12-08 09:11:43 MSK DEBUG1 cleanupThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 main: done
2009-12-08 09:11:43 MSK DEBUG1 slon: restart of worker
2009-12-08 09:11:43 MSK CONFIG main: slon version 1.2.17 starting up
2009-12-08 09:11:43 MSK CONFIG main: local node id = 2
2009-12-08 09:11:43 MSK CONFIG main: launching sched_start_mainloop
2009-12-08 09:11:43 MSK CONFIG main: loading current cluster configuration
2009-12-08 09:11:43 MSK CONFIG storeNode: no_id=1 no_comment='director'
2009-12-08 09:11:43 MSK CONFIG storePath: pa_server=1 pa_client=2
pa_conninfo="dbname=plus1 host=192.168.1.137 port=5432 user=slony"
pa_connretry=10
2009-12-08 09:11:43 MSK CONFIG storeListen: li_origin=1 li_receiver=2
li_provider=1
2009-12-08 09:11:43 MSK CONFIG storeSet: set_id=1 set_origin=1
set_comment='A replication set so boring no one thought to give it a
name'
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG storeSubscribe: sub_set=1
sub_provider=1 sub_forward='t'
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG enableSubscription: sub_set=1
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG main: configuration complete - starting threads
2009-12-08 09:11:43 MSK DEBUG1 localListenThread: thread starts
NOTICE:  Slony-I: cleanup stale sl_nodelock entry for pid=31722
2009-12-08 09:11:43 MSK CONFIG enableNode: no_id=1
2009-12-08 09:11:43 MSK DEBUG1 remoteListenThread_1: thread starts
2009-12-08 09:11:43 MSK DEBUG1 remoteWorkerThread_1: thread starts
2009-12-08 09:11:43 MSK DEBUG1 cleanupThread: thread starts
2009-12-08 09:11:43 MSK DEBUG1 syncThread: thread starts
2009-12-08 09:11:43 MSK DEBUG1 main: running scheduler mainloop
2009-12-08 09:11:43 MSK DEBUG1 remoteWorkerThread_1: helper thread for
provider 1 created
2009-12-08 09:11:43 MSK DEBUG1 remoteListenThread_1: connected to
'dbname=plus1 host=192.168.1.137 port=5432 user=slony'
2009-12-08 09:11:43 MSK INFO   Checking local node id
2009-12-08 09:11:43 MSK INFO   Found local node id
2009-12-08 09:11:43 MSK INFO   prepare for DDL script - set:1 onlyonnode:-1
2009-12-08 09:11:43 MSK ERROR  remoteWorkerThread_1: "select
"_slony".ddlScript_prepare_int(1, -1); " PGRES_FATAL_ERROR ERROR:
Slony-I: alterTableRestore(): Table "public"."access_type" is not in
altered state
CONTEXT:  SQL statement "SELECT  "_slony".alterTableRestore( $1 )"
PL/pgSQL function "ddlscript_prepare_int" line 46 at PERFORM
2009-12-08 09:11:43 MSK ERROR  remoteWorkerThread_1: DDL preparation
failed - set 1 - only on node -1
2009-12-08 09:11:43 MSK DEBUG1 slon: retry requested
2009-12-08 09:11:43 MSK INFO   remoteListenThread_1: disconnecting
from 'dbname=plus1 host=192.168.1.137 port=5432 user=slony'
2009-12-08 09:11:43 MSK DEBUG1 remoteListenThread_1: thread done
2009-12-08 09:11:43 MSK DEBUG1 localListenThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 syncThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 cleanupThread: thread done
2009-12-08 09:11:43 MSK DEBUG1 main: scheduler mainloop returned
2009-12-08 09:11:43 MSK DEBUG1 main: done
2009-12-08 09:11:43 MSK DEBUG1 slon: restart of worker
2009-12-08 09:11:43 MSK CONFIG main: slon version 1.2.17 starting up
2009-12-08 09:11:43 MSK CONFIG main: local node id = 2
2009-12-08 09:11:43 MSK CONFIG main: launching sched_start_mainloop
2009-12-08 09:11:43 MSK CONFIG main: loading current cluster configuration
2009-12-08 09:11:43 MSK CONFIG storeNode: no_id=1 no_comment='director'
2009-12-08 09:11:43 MSK CONFIG storePath: pa_server=1 pa_client=2
pa_conninfo="dbname=plus1 host=192.168.1.137 port=5432 user=slony"
pa_connretry=10
2009-12-08 09:11:43 MSK CONFIG storeListen: li_origin=1 li_receiver=2
li_provider=1
2009-12-08 09:11:43 MSK CONFIG storeSet: set_id=1 set_origin=1
set_comment='A replication set so boring no one thought to give it a
name'
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG storeSubscribe: sub_set=1
sub_provider=1 sub_forward='t'
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG enableSubscription: sub_set=1
2009-12-08 09:11:43 MSK WARN   remoteWorker_wakeup: node 1 - no worker thread
2009-12-08 09:11:43 MSK CONFIG main: configuration complete - starting threads
2009-12-08 09:11:43 MSK DEBUG1 localListenThread: thread starts
2009-12-08 09:11:43 MSK FATAL  localListenThread: "select
"_slony".cleanupNodelock(); insert into "_slony".sl_nodelock values (
  2, 0, "pg_catalog".pg_backend_pid()); " - ERROR:  duplicate key
value violates unique constraint "sl_nodelock-pkey"

2009-12-08 09:11:43 MSK DEBUG1 slon: shutdown requested
2009-12-08 09:12:03 MSK DEBUG1 slon: child termination timeout - kill child
2009-12-08 09:12:03 MSK DEBUG1 slon: done


-- 
=======================================================
Я в контакте: http://vkontakte.ru/id16323414
Сила солому ломит, но не все в нашей жизни - солома, да и сила далеко не все.