Thanks Andreas,

Yes eventually figured out I could just run a blank sql using execute script
and that appears to have fixed the problem.  I thought perhaps it only
rebuilt triggers for tables modified in the SQL or something tricky.

Thanks very much for the response, I'm interested in what you mean when you
say run it node by node.  How can one run it node by node ? I'm just passing
slonik a script such as

cluster name =3D cluster;
node 2 conninfo=3D"xxx"
node 3 conninfo=3D"xxx"
  execute script (
    set id =3D 1,
    filename =3D '/tmp/file.sql',
    event node =3D 3
  );


On Mon, Apr 7, 2008 at 4:27 PM, Andreas Kostyrka <andreas at kostyrka.org>
wrote:

> Run an EXECUTE SCRIPT on the cluster?
> The good part here is that you can run it node by node to help locking
> contention.
>
> Andreas
>
> Am Montag, den 07.04.2008, 09:44 +1000 schrieb James Tucker:
> > Hi All,
> >
> > We have recently experienced an incident whereby certain tables were
> > directly updated and new columns and constraints added without using
> > the execute script function.  We are now receiving errors (ERROR:
> > Slony-I: old key column TTT.FFF IS NULL on {UPDATE,DELETE}) on certain
> > transactions, and these transactions are being rolled back.  The DDL
> > changes were also done on the slave node, so schemas are are both
> > correct.  I believe the issue is to do with the triggers not being
> > updated after .
> >
> > What is the best way to update/rebuild the existing triggers on tables
> > which have been modified ?
> >
> > Thanks
> >
> > James
> > _______________________________________________
> > Slony1-general mailing list
> > Slony1-general at lists.slony.info
> > http://lists.slony.info/mailman/listinfo/slony1-general
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20080407/=
cd51fc06/attachment.htm