From owner-freebsd-hackers Thu Jan 23 11:59:11 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id LAA08322 for hackers-outgoing; Thu, 23 Jan 1997 11:59:11 -0800 (PST) Received: from thelab.hub.org (hal-ns1-39.netcom.ca [207.181.94.103]) by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id LAA08311 for ; Thu, 23 Jan 1997 11:59:06 -0800 (PST) Received: from thelab.hub.org (LOCALHOST [127.0.0.1]) by thelab.hub.org (8.8.4/8.8.2) with SMTP id PAA15424 for ; Thu, 23 Jan 1997 15:58:57 -0400 (AST) Date: Thu, 23 Jan 1997 15:58:57 -0400 (AST) From: The Hermit Hacker To: hackers@freebsd.org Subject: open up your (TCP) window (fwd) Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hopefully without starting a major debate *grin*...can anyone with a firm grasp of what this guy is asking for explain it to me? About the only thing I can understand from this is that it seems that the FreeBSD box is isn't reading from the socket fast enough, or in large enough chunks? ---------- Forwarded message ---------- Date: Thu, 23 Jan 1997 05:48:31 -0800 (PST) To: scrappy@hub.org Subject: open up your (TCP) window I pointed etherfind at a TCP connection between our servers and watched the transfer of a large article to your server. > Using interface le0 > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9CE08, window 799, <8> <10> Your server tells mine that it can accept 799 bytes of data. > TCP from news.gv.tsc.tdk.com.3551 to hub.org.nntp seq 66D9CE08, ack 510D1447, window 17280, <8> <10> 799 bytes data Mine sends it. > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9D127, window 0, <8> <10> Your server tells my server that it got the data and isn't ready for any more right now. > TCP from news.gv.tsc.tdk.com.3551 to hub.org.nntp seq 66D9D127, ack 510D1447, window 17280, <8> <10> 1 bytes data My server got impatient because it was in this state for a while and thought the packet from your server telling my server that it it could send more got lost on the net. So my server sent yours one byte of data (a zero window probe). > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9D128, window 0, <8> <10> Your server says it's still not ready for more. > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9D128, window 3071, <8> <10> Finally your server says send 3071 bytes. Even though it's TCP buffers were chock full of data for this connection (whatever your TCP receive buffer size is set to), innd only drained 3071 bytes when it got around to reading on the fd for this connection. What's your TCP receive buffer size set to? > TCP from news..gv.tsc.tdk.com.3551 to hub.org.nntp seq 66D9D128, ack 510D1447, window 17280, <8> <10> 1440 bytes data > TCP from news.gv.tsc.tdk.com.3551 to hub.org.nntp seq 66D9D6C8, ack 510D1447, window 17280, <8> <10> 1440 bytes data My server sends two big (maximal size?) packets in a row. > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9DC68, window 191, <8> <10> Your server acknowledges receipt, and says it still has 191 bytes of room. Apparently innd hasn't drained the buffers in the meantime (3071-1440-1440 == 191). > TCP from news.gv.tsc.tdk.com.3551 to hub.org.nntp seq 66D9DC68, ack 510D1447, window 17280, <8> <10> 191 bytes data My server sends the remaining amount. > TCP from hub.org.nntp to news.gv.tsc.tdk.com.3551 seq 510D1447, ack 66D9DD27, window 0, <8> <10> Your server got the data and isn't yet ready for more. --- Truck