From owner-freebsd-questions Mon Jan 21 5:21: 1 2002 Delivered-To: freebsd-questions@freebsd.org Received: from genesis.powertrip.net (genesis.powertrip.net [204.216.129.4]) by hub.freebsd.org (Postfix) with ESMTP id 084CF37B404 for ; Mon, 21 Jan 2002 05:20:55 -0800 (PST) Received: from localhost (epiphani@localhost) by genesis.powertrip.net (8.11.1/8.9.3) with ESMTP id g0LDKtl15569 for ; Mon, 21 Jan 2002 05:20:55 -0800 Date: Mon, 21 Jan 2002 05:20:55 -0800 (PST) From: Aaron To: freebsd-questions@freebsd.org Subject: Odd webserver problem.. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Greetings, I'm cross-posting this with the freebsd support people and an apache support list, as I'm downright confused as to what is causing my problem. I administrate a number of machines, one of which doing primarily virtual web hosting on a FreeBSD 4.4-REL machine using apache 1.3.22. The throughput is often peeks slightly over 15 Mbits, all of that being http traffic. However, at seemingly random times, the machine goes completely unreachable, or becomes extremely slow in all ip socket operations. Sometimes this can happen for as little as 30 minutes, other times can last for in excess of 6 hours (if it goes only slow, where icmp and tcp traffic still get through, but any content retrieval fails - which is why i dont get paged.) I've spend the past few weeks attempting to find some way of finding the cause of this, but I'm stumped. When the machine goes completely unreachable, it comes back by itself, without having rebooted. When I find the server laggy, after waiting 5 minutes to log in, i restart apache, and it kicks back up. The load averages during this go through the roof, but process number, cpu usage, ram usege all seem normal. So heres my problem. If it is an apache problem, why does the entire machine respond as if im getting hit with 40 megs of UDP. (which im not, already checked that.) If its a socket handling problem, why does an apache restart fix it. Here are some machine specs: Dual Xeon 600 w/ 512ram && 1 gig swap dual 18 gig scsi && 30 gig ide FreeBSD 4.4-REL pseudo-device bpf #Berkeley packet filter options ICMP_BANDLIM #Rate limit bad replies options KTRACE #ktrace(1) support kern.maxfiles: 16424 kern.maxvnodes: 40211 kern.maxproc: 8212 kern.maxfilesperproc: 16424 kern.maxprocperuid: 8211 kern.ipc.nmbclusters: 32768 Some Apache info: ServerType standalone Timeout 300 KeepAlive On MaxKeepAliveRequests 100 KeepAliveTimeout 15 MinSpareServers 100 MaxSpareServers 200 StartServers 600 MaxClients 4096 MaxRequestsPerChild 30 (Other than the paths and fd table, its a default build for freebsd, with all the modules.) Any help anyone can give would be very much appreciated. I'm not a member of either of these lists, so please remember to include my address in responces. Thanks, Aaron Wiebe Quik-Host.net Systems Admin DALnet Senior Coder To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message