From owner-freebsd-questions Mon Aug 3 05:40:47 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id FAA18639 for freebsd-questions-outgoing; Mon, 3 Aug 1998 05:40:47 -0700 (PDT) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from www.schell.de ([195.20.238.2]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id FAA18632 for ; Mon, 3 Aug 1998 05:40:34 -0700 (PDT) (envelope-from sas@schell.de) Received: from localhost (sas@localhost [127.0.0.1]) by www.schell.de (8.9.0/8.9.0) with SMTP id OAA12392; Mon, 3 Aug 1998 14:40:17 +0200 Date: Mon, 3 Aug 1998 14:40:17 +0200 (MET DST) From: Sascha Schumann To: Brian Tiemann cc: freebsd-questions@FreeBSD.ORG Subject: Re: More httpd process-limit problems In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Maybe this problem is related to the FIN_WAIT_2 problem encountered with older releases of Apache. I'm no TCP expert and don't have the RFC in mind, but if you run a newer Apache release (e.g. 1.3.1) then there shouldn't be a problem. http://www.apache.org/docs/misc/fin_wait_2.html discusses the problem. You didn't write, if your server is under heavy load or if the problem occurs after many hits. Anything unusual in the logs? Greetings, Sascha On Mon, 3 Aug 1998, Brian Tiemann wrote: > > Okay-- so I've been able to fix the first part of my problem, > thanks to the advice on this list-- that part was getting apache to spawn > more than 128 or so total processes. However, I've got a new phenomenon > happening; I hope someone can enlighten me: > > Most of the time, my server sits at about 60 processes, 30 of > which are httpd. However, every so often (every two days or so), the > number of httpd processes will suddenly just start building up; old > processes won't close, and I'll look and find 200 or 300 httpd processes > sitting idle and not allowing any new ones to spawn. > > The output of netstat shows that most of the dead connections are > in the TIME_WAIT state; restarting httpd will kill the dead processes > (they go zombie, and then the process leader kills them on the second or > third round of kill signals), but the connections remain open in > TIME_WAIT. Apache's server-info handler page shows that all the dead > connections are still in the "W" (write) phase. > > What's going on here? I don't want to have to keep running top to > see whether the server's going haywire every few minutes. Here's my > config, again, for the record: > > kern.maxvnodes: 11907 > kern.maxproc: 4116 > kern.maxfiles: 8232 > kern.argmax: 65536 > kern.securelevel: -1 > kern.hostid: 0 > kern.clockrate: { hz = 100, tick = 10000, profhz = 1024, stathz = 128 } > kern.posix1version: 199009 > kern.ngroups: 16 > kern.job_control: 1 > kern.saved_ids: 0 > kern.boottime: { sec = 900759813, usec = 736527 } Sat Jul 18 04:03:33 1998 > kern.domainname: > kern.update: 30 > kern.osreldate: 226000 > kern.bootfile: /kernel > kern.maxfilesperproc: 8232 > kern.maxprocperuid: 4115 > kern.dumpdev: { major = 255, minor = -65281 } > kern.somaxconn: 256 > kern.maxsockbuf: 262144 > kern.ps_strings: -272637968 > kern.usrstack: -272637952 > kern.shutdown_timeout: 120 > kern.acct_suspend: 2 > kern.acct_resume: 4 > kern.acct_chkfreq: 15 > kern.quantum: 10 > kern.sockbuf_waste_factor: 8 > kern.consmute: 0 > > And, in login.conf: > > www:\ > :path=/bin /sbin /usr/bin /usr/sbin /usr/local/bin /usr/local/sbin:\ > :cputime=infinity:\ > :filesize=128M:\ > :datasize-cur=64M:\ > :stacksize-cur=32M:\ > :coredumpsize-cur=0:\ > :maxmemorysize-cur=128M:\ > :memorylocked=32M:\ > :maxproc=512:\ > :openfiles=512:\ > :tc=default: > > > Can anyone offer some suggestions? Thanks very much! > > Brian > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-questions" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message