[Pgpool-general] Idea: peer-to-peer app architecture with pgpool?

[Jaume Sabater]

On Thu, Feb 5, 2009 at 12:37 AM, alex <koalillo at fastmail.fm> wrote:

> I'm thinking about setting up a cluster architecture for a web app. I
> was thinking about running an "everything replicated on all boxes" arch.
>
> That is, I run a couple of nodes, each with a web/app server + pgpool +
> PostgreSQL. pgpool set to replicate write transactions to all nodes, but
> run read queries in the local node.
>
> This way, read-only queries do not have network overhead, yet they are
> load-balanced. I can have a simple high-availability setup (just shut
> off nodes).
>
> Is this a terrible idea? Thoughts?

More details are needed to give a proper answer but, from a first
impression, no, it is not. It's quite common, actually.

-- 
Jaume Sabater
http://linuxsilo.net/

"Ubi sapientas ibi libertas"