Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 2 Jun 1999 14:32:40 -0700
From:      "David Schwartz" <davids@webmaster.com>
To:        "Matthew Hunt" <mph@astro.caltech.edu>, "Andre Oppermann" <oppermann@pipeline.ch>
Cc:        <chat@freebsd.org>
Subject:   RE: net.inet.tcp.always_keepalive on as default ?
Message-ID:  <000001bead3f$711c0840$021d85d1@whenever.youwant.to>
In-Reply-To: <19990602140108.A47013@wopr.caltech.edu>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help

> That's how keepalives work.  My understanding is that David Schwartz's
> comment referred to application idle timeouts, not keepalives.

	Correct. Only the application can know whether the most logical means of
dealing with the TCP connection is by a keepalive, data timeout, or other
rule. Applications that have no data timeouts and do not enable keepalives
are broken, period.

	I like the idea of having a ridiculously high keepalive time (like 1 week)
even for connections that don't request them, and reverting to normal
keepalive times for applications that do. That will help workaround broken
applications.

	DS

	PS: Moved to chat since this is starting to look more like a holy war. ;)



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message




Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?000001bead3f$711c0840$021d85d1>