From owner-freebsd-stable@FreeBSD.ORG Tue Jan 31 15:14:10 2006 Return-Path: X-Original-To: freebsd-stable@freebsd.org Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4385A16A422 for ; Tue, 31 Jan 2006 15:14:10 +0000 (GMT) (envelope-from holtor@yahoo.com) Received: from web31711.mail.mud.yahoo.com (web31711.mail.mud.yahoo.com [68.142.201.191]) by mx1.FreeBSD.org (Postfix) with SMTP id 0456E43D64 for ; Tue, 31 Jan 2006 15:13:49 +0000 (GMT) (envelope-from holtor@yahoo.com) Received: (qmail 55022 invoked by uid 60001); 31 Jan 2006 15:13:49 -0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=4NQuNIf4mGTNg0306bgSr4thrxpDt3tUyGuy9v28OURDZeCG7WFAf9oYCB08iJEzSt7EGVANj/h/5+FV8/2xZvPwzni2kNNPhi1shbn2BNWD0CWuxlqoWs/fUXwERqHyjMj5jbUjzYhFtVq67XtAAO/QTKn7PcY+eLeIy9Frwx4= ; Message-ID: <20060131151349.55020.qmail@web31711.mail.mud.yahoo.com> Received: from [67.85.37.167] by web31711.mail.mud.yahoo.com via HTTP; Tue, 31 Jan 2006 07:13:49 PST Date: Tue, 31 Jan 2006 07:13:49 -0800 (PST) From: Holtor To: freebsd-stable@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Subject: Network Connection Freezing X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 31 Jan 2006 15:14:10 -0000 Hello, We have a fairly busy web server runnning 5.4-STABLE and the ethernet connection seems to drop offline during the peak times daily. We have seen errors such as this: Jan 29 17:06:00 web kernel: sis0: initialization failed: no memory for rx buffers Jan 29 17:13:01 web kernel: sis0: initialization failed: no memory for rx buffers Logging into the console and doing an ifconfig sis0 down and then up would bring the server back online. The FreeBSD OS never crashed. I saw other sis0 cable fix errors so yesterday I removed this sis0 card and replaced it with an Intel card (fxp0). Much to my dismay the network connection died again this morning. The fxp0 card reported no errors on the console however when I did the usual ifconfig down then back up this error came up on the console: Jan 31 17:51:10 web kernel: in_cksum_skip: out of data by 15320 After the first set of errors I modified certain sysctl variables: kern.maxfiles=32768 kern.maxfilesperproc=16384 kern.ipc.shmall=16384 kern.ipc.nmbclusters=65535 kern.ipc.somaxconn=8192 net.inet.ip.dummynet.hash_size=2048 net.inet.ip.fw.one_pass=1 net.inet.ip.rtexpire=2 net.inet.ip.rtminexpire=2 net.inet.ip.redirect=0 net.inet.tcp.delayed_ack=0 This web server is a music download server which is has a varying ipfw per destination IP download limit depending on the time of day to control bandwidth usage. This causes httpd to run 500, 600, 700 times. RAM is not an issue since we upgraded to 2 GB. Does anyone have any idea how to solve this problem? Perhaps there's other sysctl variables we can increase? Any kernel options? Thank you very much in advance! Holt G. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com