From owner-freebsd-stable Thu Jul 11 7:21:33 2002 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 8110137B400 for ; Thu, 11 Jul 2002 07:21:28 -0700 (PDT) Received: from russian-caravan.cloud9.net (russian-caravan.cloud9.net [168.100.1.4]) by mx1.FreeBSD.org (Postfix) with ESMTP id 32FA043E09 for ; Thu, 11 Jul 2002 07:21:28 -0700 (PDT) (envelope-from mark@cloud9.net) Received: from earl-grey.cloud9.net (earl-grey.cloud9.net [168.100.1.1]) by russian-caravan.cloud9.net (Postfix) with ESMTP id B1F7628D7D for ; Thu, 11 Jul 2002 10:21:27 -0400 (EDT) Date: Thu, 11 Jul 2002 10:21:27 -0400 (EDT) From: Mark Hennessy To: Subject: Re: Cistron 1.6.6 freeze-ups (fwd) Message-ID: <20020711101547.A89350-100000@earl-grey.cloud9.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Hi all. I'm using FreeBSD 4.5 RELEASE, and I was wondering if any of the following message makes sense. Basically, I'm running Cistron RADIUS 1.6.6 among other things on my FreeBSD box and as of late have run into a problem where the authentication process seems to die for no apparent reason. This has been happening every few hours for a few days now and I was wondering if there was any advice on how to further troubleshoot/resolve the problem. The only way to bring the process back to life is to kill radiusd off and restart it fresh. They seem to think that it's a UDP problem in FreeBSD but have no other ideas. Any comments or ideas would be appreciated. -- Mark P. Hennessy mark@cloud9.net ---------- Forwarded message ---------- Date: Tue, 9 Jul 2002 18:55:44 +0200 (CEST) From: Emile van Bergen Reply-To: cistron-radius@lists.cistron.nl To: "cistron-radius@lists.cistron.nl" Subject: Re: Cistron 1.6.6 freeze-ups On Tue, 9 Jul 2002, Mark Hennessy wrote: > > > Still nothing from ktrace.out (strace doesn't compile, so I use the > > > included FreeBSD ktrace) > > > > What does ps show as the syscall the processes are sleeping on (look at > > the 'wchan' column)? What's the process state? Are they STOPed, for > > whatever reason, eg. because of a failed trace attempt (state T)? > > > UID PID PPID CPU PRI NI VSZ RSS WCHAN STAT TT TIME COMMAND > 0 25033 1 0 2 0 1904 1052 sbwait Is ?? 0:10.64 /usr/local/sbin/radiusd > 0 25034 25033 0 2 0 1776 996 select S ?? 0:00.68 /usr/local/sbin/radiusd Hmm, it's in sbwait, which seems to be a kernel routine that is called when recvfrom() blocks. But if the server would block there, it would become runnable again as soon as something comes in (which does happen, see below), and that doesn't happen. I don't know if it's related, but there is a bug in NetBSD (you use FreeBSD, I know) that had something to do with the kernel looping in sbwait, look at http://www.geocrawler.com/archives/3/495/2000/5/50/3799097/ > The server hasn't been bound to any specific outside interface at the > commandline. > > netstat -a | head -2 ; netstat -a | grep rad > > Active Internet connections (including servers) > Proto Recv-Q Send-Q Local Address Foreign Address (state) > udp4 0 0 *.radacct *.* > udp4 14238 0 *.radius *.* ^^^^^ This is an interesting number. Apparently the kernel receives the packets allright, but doesn't pass them on to Cistron for whatever reason. It looks like you have a FreeBSD bug...? Cheers, Emile. -- E-Advies / Emile van Bergen | e-advies@evbergen.xs4all.nl tel. +31 (0)70 3906153 | http://www.e-advies.info - List info/subscribe/unsubscribe? See http://www.radius.cistron.nl/list/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message