On Fri, Jul 09, 2010 at 01:25:04PM -0700, Waldo Nell wrote:
> but table XXX contains approximately 130 million rows.  This will translate to 130 million update statements via Slony-I to the other replicated node.  Any more efficient way to update this?  If not, will Slony-I break on this update?
> 

It probably won't break, but I bet you won't be happy.

When I used to have to do this sort of thing, I'd do it in transaction
batches of a few hundred or thousand at a time.  If you can't do it
that way, the only real effective way to do it is to drop that table
from replication, make the changes, and then add the table again.

A

-- 
Andrew Sullivan
ajs at crankycanuck.ca