Re: Scaling Database for heavy load

From: Chris Travers <chris(dot)travers(at)gmail(dot)com>
To: Digit Penguin <digitpenguin(at)gmail(dot)com>
Cc: Postgres General <pgsql-general(at)postgresql(dot)org>
Subject: Re: Scaling Database for heavy load
Date: 2016-05-11 10:25:41
Message-ID: CAKt_ZfsOni=JJiL=b=Faj9doyQN2mfM12pijvxfxc-d+i07sbw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Wed, May 11, 2016 at 12:09 PM, Digit Penguin <digitpenguin(at)gmail(dot)com>
wrote:

> Hello,
>
>
> we use PostgreSql 9.x in conjunction with BIND/DNS for some Companies with
> about 1.000 queries per second.
> Now we have to scale the system up to 100.000 queries per second (about).
>
> Bind/DNS is very light and i think can not give us bottleneck.
> The question is how to dimension the backend database.
>
> The queries are select (only few insert or update), but the 100.000
> queries per second are only select.
>
> How can i calculate/dimensionate?
> We think to put mor ethan one Bind Server (with backend database) behinbd
> a router with balancing capabilities.
>
> The problem is to know which requirements and limits does a Postgresql 9.x
> installation - 64 bit - can have.
> Furthermore, we tried Rubyrep (it is quite old!); can you suggest me other
> replication modules that can work also if connction link, from Database
> Server, went down?
>

If they are almost all select queries and a little lag between write and
read visibility is ok, I would recommend Streaming replication, Slony, or
Bucardo and to query against your replicas. A specific architecture using
one or more of these replication technologies would need to be designed
based on your specific needs of course.

>
> Thank you!
> Francesco
>

--
Best Wishes,
Chris Travers

Efficito: Hosted Accounting and ERP. Robust and Flexible. No vendor
lock-in.
http://www.efficito.com/learn_more

In response to

Browse pgsql-general by date

  From Date Subject
Next Message martin.kamp.jensen 2016-05-11 10:44:03 Re: Invalid data read from synchronously replicated hot standby
Previous Message Digit Penguin 2016-05-11 10:09:25 Scaling Database for heavy load