I couldn't find anything on pgpool regarding sharding, but I did stumble upon this
presentation that mentioned pl/proxy. PL/Proxy looks good for sharding, though it looks like it could really limit the ability to do ad hoc queries. But I suppose that would work for OLTP and then we could just maintain a separate OLAP database that may have to read from disk but would have the advantage of existing on a single machine....
-Alessandro
On Fri, Jun 15, 2012 at 12:17 PM,
ktm@rice.edu <ktm@rice.edu> wrote:
On Fri, Jun 15, 2012 at 11:58:02AM -0700, Alessandro Gagliardi wrote:
> I pretty much consider anything v1.0 to be "beta" for all intents and purposes.
>
> That said, perhaps the intersection between "web scale" and "tried and
> true" is a null set. I hope it's not. I guess what I'm hoping for is a
> methodology that uses tried and true technology (as opposed to untested
> bleeding edge technology).
>
You may want to look at pgpool to shard across a set of databases.
Good luck and keep us updated with what you do and how it worked out.
Regards,
Ken