After heroku's most recent outage, I've been trying to find a way to replicate my heroku database to a non-heroku server and so far haven't found an easy way to do this. Has anybody gotten the log streaming to work with a postgres box that is hosted somewhere else (in my case on Rackspace)?

A question : is there any way for my application code to transparently "recognize" that (at runtime) there is a new follower PG database that is assigned to the application and that it should route some of the reads to the new follower DB? Or do I have to manually change my app's db configuration and include the new follower's address?

Does Postgres actually natively support this and is there any client library that makes use of that? I know that Zookeeper serves this sort of purpose (automatic service discovery etc.), but I have no idea whether/how it works with Postgres.

Can someone more versed in the current state of Postgres chime in on something for me?

> "One use case that has historically been challenging in database management is setting up a read replica, often referred to as a read slave."

Having only inherited MySQL stuff in production, this is pretty trivial in MySQL land - I was under the impression that Postgres also had a solution for this, with Slony and recently as a part of Postgres core. Is my info wrong on this?

I thought master/slave replication was a pretty solved problem at this point, am I wrong?

Can anyone comment on how this affects consistency from a writing client?

For instance, let's say I update a user record with a new email address, then display then redirect to an action to display the user record. The write may not have reached the replica, so I may get the old email address, correct? It seems this might require quite a bit more thought than simply enabling a replica.

This is nice, but it would be even better if we could provision a follower in another AWS Region or at least another AZ.

this is just so awesome... I have become quite well versed in setting up streaming replication by myself (postgres 9.1) and have a good system when i need to provision one but a one click solution is a whole other level!

I can't wait to give them my bucks, once they become Safe Harbor compliant!

Are their Postgresql offering running on top of EBS instances ?