DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=mime-version:in-reply-to:references:date:message-id:subject:from:to
	:cc:content-type:content-transfer-encoding;
	b=HFFBqsPaKbZOWroB79K6C37d+kApYS41Q59gm6ha09XCTAi3ghcQTXf9E0SBLhDL1A
	wrVeAeOP6vr51MfwAOtFYc4b+fagpNln7PKCtKW8fv3sRvhhT90lvxa9n5W3QJsH9Zto
	uEx40RamNaB+keu1gg8wNu/ICZPS5hqnLbTfc=
MIME-Version: 1.0
In-Reply-To: <1238141231.16568.574.camel@ebony.2ndQuadrant>
References: <49C9F49B.8020003@oss.ntt.co.jp>
	<1238091273.16568.530.camel@ebony.2ndQuadrant>
	<49CC75E3.7010101@oss.ntt.co.jp>
	<1238139418.16568.555.camel@ebony.2ndQuadrant>
	<49CC8524.4090103@oss.ntt.co.jp>
	<1238141231.16568.574.camel@ebony.2ndQuadrant>
Date: Fri, 27 Mar 2009 10:57:44 +0100
Message-ID: <1d4e0c10903270257v5f0f2fdev6b742066e13fd974@mail.gmail.com>
Subject: Re: display previous query string of idle-in-transaction
From: Guillaume Smet <guillaume.smet@gmail.com>
To: Simon Riggs <simon@2ndquadrant.com>
Cc: Tatsuhito Kasahara <kasahara.tatsuhito@oss.ntt.co.jp>,
	pgsql-hackers@postgresql.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

On Fri, Mar 27, 2009 at 9:07 AM, Simon Riggs <simon@2ndquadrant.com> wrote:
> Or take it further back still and think about why "idle in transaction"
> occurs at all and fix *that*. Maybe not in Postgres at all, possibly in
> the driver or even higher up client stack.

From my experience, the main reasons for this problem are:
- a poorly coded connection pool (a lot of people think they really
should write their connection pooling solution themselves...);
- a poorly coded application;
- bugs in the code.

So you really should try to fix them in your application first.

That said, I have one case in mind where I wasn't able to fix
completely the connection pool and the application and we still
encounter idle in transaction connections from time to time.

For this sort of case, a timeout would be a nice solution.

-- 
Guillaume