Re: Parallel queries for a web-application |performance testing

From: Matthew Wakeling <matthew(at)flymine(dot)org>
To: Balkrishna Sharma <b_ki(at)hotmail(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: Parallel queries for a web-application |performance testing
Date: 2010-06-17 09:41:44
Message-ID: alpine.DEB.2.00.1006171032120.2534@aragorn.flymine.org
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-performance

On Wed, 16 Jun 2010, Balkrishna Sharma wrote:
> Hello,I will have a web application having postgres 8.4+ as backend. At
> any given time, there will be max of 1000 parallel web-users interacting
> with the database (read/write)I wish to do performance testing of 1000
> simultaneous read/write to the database.

When you set up a server that has high throughput requirements, the last
thing you want to do is use it in a manner that cripples its throughput.
Don't try and have 1000 parallel Postgres backends - it will process those
queries slower than the optimal setup. You should aim to have
approximately ((2 * cpu core count) + effective spindle count) number of
backends, as that is the point at which throughput is the greatest. You
can use pgbouncer to achieve this.

> I can do a simple unix script on the postgres server and have parallel
> updates fired for example with an ampersand at the end. Example:

> echo '\timing \\update "DAPP".emp_data set f1 = 123where emp_id =0;' |
> "psql" test1 postgres|grep "Time:"|cut -d' ' -f2- >>
> "/home/user/Documents/temp/logs/$NUM.txt" &pid1=$! echo '\timing
> \\update "DAPP".emp_data set f1 = 123 where emp_id =2;' | "psql" test1
> postgres|grep "Time:"|cut -d' ' -f2- >>
> "/home/user/Documents/temp/logs/$NUM.txt" &pid2=$! echo '\timing
> \\update "DAPP".emp_data set f1 = 123 where emp_id =4;' | "psql" test1
> postgres|grep "Time:"|cut -d' ' -f2- >>
> "/home/user/Documents/temp/logs/$NUM.txt" &pid3=$! .............

Don't do that. The overhead of starting up an echo, a psql, and a grep
will limit the rate at which these queries can be fired at Postgres, and
consume quite a lot of CPU. Use a proper benchmarking tool, possibly on a
different server.

Also, you should be using a different username to "postgres" - that one is
kind of reserved for superuser operations.

Matthew

--
People who love sausages, respect the law, and work with IT standards
shouldn't watch any of them being made. -- Peter Gutmann

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Pierre C 2010-06-17 11:12:22 Re: Parallel queries for a web-application |performance testing
Previous Message venu madhav 2010-06-17 09:08:05 Obtaining the exact size of the database.