From owner-freebsd-hackers Thu Feb 25 21:57:17 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from cygnus.rush.net (cygnus.rush.net [209.45.245.133]) by hub.freebsd.org (Postfix) with ESMTP id 298BE14EB0 for ; Thu, 25 Feb 1999 21:57:15 -0800 (PST) (envelope-from bright@cygnus.rush.net) Received: from localhost (bright@localhost) by cygnus.rush.net (8.9.3/8.9.3) with SMTP id AAA00994; Fri, 26 Feb 1999 00:56:49 -0500 (EST) Date: Fri, 26 Feb 1999 00:56:47 -0500 (EST) From: Alfred Perlstein To: Dave Hayes Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Quake2 and UDP recvspace In-Reply-To: <199902260437.UAA09666@hokkshideh.jetcafe.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Thu, 25 Feb 1999, Dave Hayes wrote: > I'm runing Quake2 on FreeBSD 2.2.8-RELEASE as a dedicated server. > Quake2 with many players basically slams the UDP layer of my > machine, for those of you who don't already know. :) > > Netstat -s reports: > > udp: > 110301027 datagrams received > 0 with incomplete header > 30 with bad data length field > 1377 with bad checksum > >> 29882 dropped due to no socket > 651167 broadcast/multicast datagrams dropped due to no socket > >> 15494 dropped due to full socket buffers > 0 not for hashed pcb > 109603077 delivered > 36475854 datagrams output Just from what they mean, i would ignore "29882 dropped due to no socket" that means people were sending you UDP on ports there was no listener. so, realistically: (15494*100) / 109603077 == 0.014136464435 0.01% Is that acceptable packet loss? I also think that it's less the kernel's fault and more the quake server's fault for not eating packets that are sent to it. basically, 0.01% of the time the quake server doesn't recv() fast enough and drops some packets. you could also imagine that some other program may not be recv() fast enough to get packets as this is _total_ not per-socket. i would suggest not worrying about it. -Alfred > > Now I have my sysctl variables set up like so: > > net.inet.udp.checksum: 1 > net.inet.udp.maxdgram: 9216 > net.inet.udp.recvspace: 246723 > > Note that the recvspace is basically as big as I can make it. I > noticed a decrease in the "dropped due to full socket buffers" when I > increased the recvspace from the default 16K value. > > I know that the max value is in the kernel somewhere. My questions > are "where?" and "what will this screw up if I change it?". > ------ > Dave Hayes - Consultant - Altadena CA, USA - dave@jetcafe.org > Keep Usenet Free - http://www.jetcafe.org/~dave/usenet > >>> The opinions expressed above are entirely my own <<< > > For hatred can never put an end to hatred, love alone can. This is an > unalterable law. -The Dhammapada > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-hackers" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message