From owner-freebsd-bugs@FreeBSD.ORG Tue Jul 27 09:38:00 2004 Return-Path: Delivered-To: freebsd-bugs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7A2AA16A4CE for ; Tue, 27 Jul 2004 09:38:00 +0000 (GMT) Received: from mail.telsatgp.com.pl (pa79.pleszew.sdi.tpnet.pl [217.96.180.79]) by mx1.FreeBSD.org (Postfix) with SMTP id 58E2F43D55 for ; Tue, 27 Jul 2004 09:37:58 +0000 (GMT) (envelope-from sgp@telsatgp.com.pl) Received: (qmail 1634 invoked from network); 27 Jul 2004 09:39:48 -0000 Received: from slawek.telsatgp.com.pl (HELO Slawek) (sgp@192.168.5.5) by pa79.pleszew.sdi.tpnet.pl with SMTP; 27 Jul 2004 09:39:48 -0000 Message-ID: <000d01c473bd$687e5720$0505a8c0@Slawek> From: "Slawek" To: Date: Tue, 27 Jul 2004 11:38:01 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1437 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1441 FL-Build: Fidolook 2002 (SL) 6.0.2800.85 - 28/1/2003 19:07:30 X-Organisation: Telsat GP Subject: network hangs after about 8 hours of uptime X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 27 Jul 2004 09:38:00 -0000 Hello! I used to use very old Pentium 100 with Free BSD 4.x as a gateway for our home network and decided a few days ago to switch to Pentium Celeron system (worked under Windows before without any issues) with Free BSD 5 (current). The main reason behind this change was my plan of experimenting with MONO in near future (Linux / Unix .NET Framework implementation) and the old machine had only 32 MB of RAM. Unfortunatelly it looks like my new Free BSD installation "hangs" just about 8 hours after boot each time. Please note that this is not a hard hang, but I'll mention what this computer has installed first: - Pentium Celerom 500 MHz - 512 MB RAM - VIA chipset - two built-in RS ports - one used for modem and the other for smart UPS - PCI hardware modem used as fax (using hylafax) - two Ethernet cards (Realtek) - new 120 GB HDD (ATA) - very old CD-ROM - an old AGP video card, I don't have normally monitor attached so I prefer to do any required debugging remotely if possible Software running on the machine: - sshd - qmail - apache 2 (with php 4) - inetd (smtp, pop3, ftp) - squid - named 9 - linux binary compatibility - ppp daemon (ppp_enable in rc.conf) Each time when about 8 hours from boot passes I cannot connect to the machine anymore, but if I were already connected by SSH, the connection stays active and I can issue commands (I can't open any incoming nor outgoing TCP nor UDP connections). If I try to interactively open outside connection (or use some domain name which would open outgoing UDP connection), the command hard hangs - ^C doesn't break it - even for ping: I can ping by IP, but trying to ping by domain name totally hangs ping. Cron jobs stops working after the hang (there are no log entries saying "cron" has launched "atrun"). Establishing new TCP connections works, but no data can be transmitted in either direction. If I configure firewall to "log and block" some packets from internal machines I can try to send those packets after the hang and those are logged. I have disabled firewall just in case this were some problem with stateful firewalling - but still hangs. I have disabled USB, still hangs. Oh, and something more - if I issue a "reboot" command after the "hang", from already opened SSH connection, the SSH connection is not closed, but just hangs, and reboot is *not* done (but fsck is not launched after hard reboot, so is looks like disk is synced). I haven't found anything interesing logged after the hang (the only suspicious thing is lack of "atrun" launched entries). Any ideas? Any configurartion files / logs should I attach? thanks for any help, Slawek Piotrowski (I've scheduled "reboot" every 6 hours as a workaround, but I'd prefer to fix it somehow)