[pgpool-general: 4315] Re: Parallel Query mode & data partitioning

[Pablo Sanchez]

[ Comments below, in-line ]


On 01/08/2016 09:36 AM, Pavels Gurskis wrote:
> Hi,

Hi Paul,

> [ Inquiry about data partitioning ]

IMO, where pgpool-II really shines is being able to distribute
read-only queries across a Streaming Replication set up.  For example,
A Master database, using synchronous Streaming Replication to keep a
Slave up to date.  I'm not sure how you plan on leveraging pgpool-II.

For example, let's say you have a single PG Instance.  Initially
you'll set up the DB server with fast and slow storage.  You'll create
your partitions on the respective storage.

Next, you'll populate the data.  Supposing you have three months
total, your initial partition definitions will place the data in the
respective partitions:  slow vs fast storage.

After one month, you'll want to manually move the previous month over
to the slower storage and place new data into the fast storage.

I think the above handles what you want to do.  But only with a single
PG Instance.

If you set up synchronous Streaming Replication, you'll need to
replicate the above in the Slave.  Once done, I think it'd be worth
introducing pgpool-II to Load Balance read-only queries.

In short, I don't think you need Parallel Query and/or pgpool-II's
data partitioning capabilities.  :)  But perhaps I'm wrong.

Anyway, the above are my thoughts .... a bit of a brainstorming.
--
Pablo Sanchez - Blueoak Database Engineering, Inc
Ph:    819.459.1926         Blog:  http://pablo-blog.blueoakdb.com
iNum:  883.5100.0990.1054