In response to Geoffrey <lists at serioustechnology.com>:

> I've noted that our restore procedure is flawed now that we've started 
> replicating our data.  Apparently our backup procedure picks up a schema 
> in our primary node that was created by slony.  Thus, when I attempted 
> to restore into a database that had a schema created prior to our slony 
> implementation, I received an error because the backup data contained 
> the slony created schema and the empty database I was attempting to load 
> did not.
> 
> Question is, what is the best solution here?  Should I permit the 
> restore to include the slony schema?  What do other folks do in this 
> situation?

There are different approaches possible.

Sometimes I do dump/restores to get a copy of live data in our testing
environment.  When I do that, I just do a full dump from the primary
system (including the slony schema) then restore that to testing systems,
then jump in and drop the slony schema.

For backup purposes, I use the -N option to pg_dump to exclude the Slony
schema.

-- 
Bill Moran
Collaborative Fusion Inc.
http://people.collaborativefusion.com/~wmoran/

wmoran at collaborativefusion.com
Phone: 412-422-3463x4023