Received: from sss.sss.pgh.pa.us (sss.pgh.pa.us [209.114.166.2] (may be
	forged)) by hub.org (8.9.3/8.9.3) with ESMTP id JAA48514
	for <pgsql-interfaces@postgreSQL.org>;
	Fri, 16 Jul 1999 09:48:01 -0400 (EDT)
	(envelope-from tgl@sss.pgh.pa.us)
Received: from sss.sss.pgh.pa.us (localhost [127.0.0.1])
	by sss.sss.pgh.pa.us (8.9.1/8.9.1) with ESMTP id JAA09491;
	Fri, 16 Jul 1999 09:46:45 -0400 (EDT)
To: "Ken J. Wright" <ken@ori-ind.com>
cc: pgsql-interfaces@postgreSQL.org
Subject: Re: [INTERFACES] backend connection lost 
In-reply-to: Your message of Thu, 15 Jul 1999 13:42:40 -0700
	<3.0.32.19990715134238.008395e0@ren.cncware.com>
Date: Fri, 16 Jul 1999 09:46:45 -0400
Message-ID: <9489.932132805@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"Ken J. Wright" <ken@ori-ind.com> writes:
> app is idle for a long period of time, the connection to the backend gets
> lost. Is there some kind of postmaster setting to keep alive? Other ideas?

There is not, but you could try patching the backend to set the
KEEPALIVE option on the TCP socket --- see StreamConnection() in
src/backend/libpq/pqcomm.c.

If that helps, I think it'd be reasonable to accept the patch as
an #ifdef'd chunk of code with a config.h variable to control it.
I'm a little hesitant to just drop it in unconditionally --- I don't
know whether KEEPALIVE is supported by all platforms...

			regards, tom lane