We are moving from a datacenter in Chile to one in AWS (in USA). We've already moved most of our servers but the database (PostgreSQL). As the source server is in a VMWARE environment, we took a copy and moved it to AWS. But now it's time to resynch both servers.
From our investigation and some previous questions, seems like Slony would be the best approach, but we are having big issues with it. The most important is the initial sync, were it seems to be generating "COPY" statements for all of the data, trusting on PKs to avoid duplicates, but it actually only needs a few statements, and not all of the database to resync, as we've already moved a copy to the cloud.
Then the question is: can one configure Slony to achieve a more efficient sync, thus avoiding such a huge network overhead of moving all data once again? Perhaps a just a PK list compare or something like that for each table involved?
A side note question would be: is really Slony the best approach? Please consider a 700GB database from Chile to US, with an actual 1MB max data transfer rate.