> 2008-07-02 15:54:29 EDT INFO   completed DDL script - run
> ddlScript_complete_int()
> 2008-07-02 15:54:31 EDT ERROR
> remoteWorkerThread_6: "notify "_dbname_replication_Event"; insert
> into "_dbname_replication".sl_event     (ev_origin, ev_seqno, ev_timestamp,
> ev_minxid, ev_maxxid, ev_xip, ev_type , ev_data1, ev_data2, ev_data3    )
> values ('6', '14405', '2008-06-06
> 14:43:59.361661', '5559800', '5559800', '', 'DDL_SCRIPT', '1', 'update
> features_address set address = ''* * CONFIDENTIAL RECORD * *'' where
> raw_address like ''* * CONFIDENTIAL%'';
>     ', '-1'); insert into "_dbname_replication".sl_confirm
> (con_origin, con_received, con_seqno, con_timestamp)    values (6,
> 8, '14405', now()); commit transaction;" PGRES_FATAL_ERROR ERROR:  could
> not find trigger 1649923
> 2008-07-02 15:54:31 EDT INFO   remoteListenThread_6: disconnecting
> from 'host=host6 dbname=dbname user=slony port=5432 password=********'
> 
> 
> This DDL update is failing (could not find trigger 1649923). Have any tables
> on the slave been modified directly, not using EXECUTE SCRIPT?
> 


Possibly, I'm sure we probably did something wrong... But we do know to use
execute script.  Guess I will just start again.  There's no way to repair
things like this is there?

The execute script is only for schema changes, right?  Would running
something like the following through execute script mess anything up?

update features_address set address = '* * CONFIDENTIAL RECORD * *' where
raw_address like '* * CONFIDENTIAL%';
    

Thanks,

Peter