[pgpool-general: 2993] Re: Connection pooling between non-identical clusters

Yugo Nagata nagata at sraoss.co.jp
Mon Jun 30 21:08:14 JST 2014


Hi Martijn and hackers,

There was a discussion about a load-balance tuing parameter in pgopol-hackers
list. Is this like the what you want?

[pgpool-hackers: 493] Re: [pgpool-general: 2762] Re: OTRS
http://www.sraoss.jp/pipermail/pgpool-hackers/2014-April/000493.html

If using dabase name, listing dbname in parameters for every ndoes
using general expression might be good idea.


Any other ideas and comments, hackers?

On Wed, 18 Jun 2014 15:54:12 +0200
Martijn van Oosterhout <kleptog at gmail.com> wrote:

> Hoi,
> 
> We have a setup where due to differing replication requirements we have
> some databases in cluster A and some in cluster B. We're also looking at
> connection pooling because there a lot of clients that aren't going to do
> doing very many queries.
> 
> Looking through the documentation it doesn't seem like pgpool2 handles this
> in a single process, so I have it going with two separate pgpool instances.
> However, I was wondering if it wouldn't be an idea to allow users to
> specify database names in the configuration and route to the backend that
> contains that database.
> 
> What I'm think of is an option like:
> 
> backend_dbnames0=foo,bar_*
> backend_dbnames1=baz
> 
> So if a connection comes in for database foo_2, it will be sent to backend0
> but not backend1.
> 
> Does this sound like something people would be interested it, or is it too
> niche?
> 
> Have a nice day,
> -- 
> Martijn van Oosterhout <kleptog at gmail.com> http://svana.org/kleptog/


-- 
Yugo Nagata <nagata at sraoss.co.jp>


More information about the pgpool-general mailing list