From owner-freebsd-isp Tue Jun 18 8:19:58 2002 Delivered-To: freebsd-isp@freebsd.org Received: from daemon.kr.FreeBSD.org (daemon.kr.freebsd.org [211.176.62.31]) by hub.freebsd.org (Postfix) with ESMTP id 892B937B406 for ; Tue, 18 Jun 2002 08:19:49 -0700 (PDT) Received: from gradius.wdb.co.kr (daemon [211.176.62.31]) by daemon.kr.FreeBSD.org (Postfix) with ESMTP id 525CF8F633 for ; Wed, 19 Jun 2002 00:19:49 +0900 (KST) Received: from localhost (localhost [127.0.0.1]) by gradius.wdb.co.kr (8.12.3/8.12.3) with ESMTP id g5I7Vgii037515 for ; Tue, 18 Jun 2002 16:31:43 +0900 (KST) (envelope-from cjh@kr.FreeBSD.org) Date: Tue, 18 Jun 2002 16:31:42 +0900 (KST) Message-Id: <20020618.163142.98892204.cjh@kr.FreeBSD.org> To: freebsd-isp@freebsd.org Subject: Intel PRO/1000 problem with 4.5R From: CHOI Junho Organization: Korea FreeBSD Users Gruop X-URL: http://www.kr.FreeBSD.org/~cjh X-Mailer: Mew version 3.0.55 on Emacs 21.2 / Mule 5.0 (SAKAKI) Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Hi, (I posted this on -net but no response, and some more info added...) I am maintaining two large-volume HTTP server. It is with one em0 gigabit NIC, and running apache(or thttpd) for users to download relatively big(>200k) files. I started this service using 4.5R. The server is Pentium III 1G, 1G RAM, 36G SCSI HDD). Server load was very high(>50) but I heard it's characteristic of apache server and I decided to upgrade to recent system and use thttpd instead of apache. When I tried to use thttpd, it shows the load under 1(wow!). 2 days ago I updated the system to 4.6-RELEASE. Whether or not using apache or thttpd, under some traffic(about >200Mbit/sec), it shows "watchdog timeout" repeatedly: Jun 16 11:36:18 gigatest /kernel: em0: Link is Down Jun 16 11:36:18 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:40:49 gigatest /kernel: em0: watchdog timeout -- resetting Jun 16 11:40:51 gigatest /kernel: em0: Link is Down Jun 16 11:40:51 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:40:51 gigatest /kernel: em0: Link is Down Jun 16 11:40:51 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:44:23 gigatest /kernel: em0: watchdog timeout -- resetting Jun 16 11:44:24 gigatest /kernel: em0: Link is Down Jun 16 11:44:24 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:50:28 gigatest /kernel: em0: watchdog timeout -- resetting Jun 16 11:50:29 gigatest /kernel: em0: Link is Down Jun 16 11:50:29 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:50:30 gigatest /kernel: em0: Link is Down Jun 16 11:50:30 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex Jun 16 11:50:30 gigatest /kernel: em0: Link is Down Jun 16 11:50:30 gigatest /kernel: em0: Link is up 1000 Mbps FullDuplex Jun 16 11:50:30 gigatest /kernel: em0: Link is Down Jun 16 11:50:30 gigatest /kernel: em0: Link is up 1000 Mbps Full Duplex I updated em0 driver to 1.3.8 from Intel webpage, but it no helps. After that day, I downgraded the system to 4.5R-p6. It still has a problem. Here is my sysctl.conf: kern.ipc.maxsockbuf=20480000 kern.ipc.somaxconn=8192 kern.ipc.maxsockets=60000 kern.maxfiles=65536 kern.maxfilesperproc=32768 net.inet.tcp.rfc1323=1 net.inet.tcp.delayed_ack=0 net.inet.tcp.sendspace=65536 net.inet.tcp.recvspace=16384 net.inet.udp.recvspace=16384 net.inet.udp.maxdgram=57344 net.local.stream.recvspace=65536 net.local.stream.recvspace=65536 net.local.stream.sendspace=65536 net.inet.tcp.msl=1000 net.inet.icmp.drop_redirect=1 net.inet.icmp.log_redirect=1 net.inet.ip.redirect=0 net.inet6.ip6.redirect=0 net.link.ether.inet.max_age=1200 net.inet.ip.sourceroute=0 net.inet.ip.accept_sourceroute=0 net.inet.icmp.bmcastecho=0 net.inet.icmp.maskrepl=0 and /boot/loader.conf.local: kern.ipc.nmbclusters="131072" accf_http_load="YES -- +++ Any opinions in this posting are my own and not those of my employers +++ CHOI Junho [sleeping now] [while sleeping] Korea FreeBSD Users Group Web Data Bank To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message