From mboxrd@z Thu Jan 1 00:00:00 1970 From: Satoru SATOH Subject: Re: 200 millisecond timeouts in TCP Date: Sat, 5 Jun 2010 00:02:02 +0900 Message-ID: <20100604150202.GA8268@localhost.localdomain> References: Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Ivan Novick , netdev@vger.kernel.org To: Ryousei Takano Return-path: Received: from mail-pw0-f46.google.com ([209.85.160.46]:62625 "EHLO mail-pw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752354Ab0FDPCJ (ORCPT ); Fri, 4 Jun 2010 11:02:09 -0400 Received: by pwj5 with SMTP id 5so27504pwj.19 for ; Fri, 04 Jun 2010 08:02:07 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: It's possible to tune the RTO min value (rto_min) per route since 2.6.23+. (see also the manual of iproute2, ip(8) ) - satoru On Fri, Jun 04, 2010 at 03:58:57PM +0900, Ryousei Takano wrote: > On Fri, Jun 4, 2010 at 7:37 AM, Ivan Novick wr= ote: > > Hello, > > > > Using tcpdump and systemtap I am seeing that sometimes retransmissi= on > > of data is sent after waiting 200 milliseconds. =C2=A0However somet= imes > > retransmissions happen quicker. > > > > Is there a specifc event that causes these 200 milisec delays to ki= ck > > in? =C2=A0Are those events identifiable in netstat -s output? > > > > Also do you know if the timeout numbers for TCP are configurable pa= rameters? > > > The minimum RTO value is fixed to 200 ms. It is useful to make the m= in/max > RTO values tunable. For example, reducing the minimum RTO value is e= ffective > for TCP incast problem [1]. Of course, it may occur spurious retrans= missions. >=20 > [1] Vijay Vasudevan, et al, Safe and Effective Fine-grained TCP Retra= nsmissions > for Datacenter Communication, SIGCOMM2009 >=20 > In Solaris, there are two tunable parameters: tcp_rexmit_interval_min= /max. >=20 > Do you have plan to introduce sysctl parameters like these to the Lin= ux. >=20 > Thanks, > Ryousei