From owner-freebsd-questions Sun May 14 16:13:12 2000 Delivered-To: freebsd-questions@freebsd.org Received: from camelot.bitart.com (BITart-45.BITart.com [206.103.221.45]) by hub.freebsd.org (Postfix) with SMTP id C6B1A37B640 for ; Sun, 14 May 2000 16:13:06 -0700 (PDT) (envelope-from gerti@bitart.com) Received: (qmail 5861 invoked by uid 101); 14 May 2000 23:13:09 -0000 Message-ID: <20000514231309.5860.qmail@camelot.bitart.com> Content-Type: text/plain MIME-Version: 1.0 (NeXT Mail 4.2mach v148) In-Reply-To: <20000514145641.O28383@fw.wintelcom.net> X-Nextstep-Mailer: Mail 4.2mach (Enhance 2.2p1) Received: by NeXT.Mailer (1.148) From: Gerd Knops Date: Sun, 14 May 2000 18:13:09 -0500 To: Alfred Perlstein Subject: Re: What's killing my processes? Cc: Andrew McNaughton , freebsd-questions@FreeBSD.ORG Reply-To: gerti@BITart.com References: <20000514145641.O28383@fw.wintelcom.net> Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Alfred Perlstein wrote: > * Andrew McNaughton [000514 09:38] wrote: > > > > Something seems to be knocking off various processes on my server > > with the equivalent of a 'kill -9'. This has included processes run > > from the shell as root (make, pico) and some cron jobs. I'm > > guessing there's more, but those are the ones I have a little > > information on. > > > > The only thing I can think of that might do something like this is > > resource limits, but that seems unlikely to hit pico in mid use on > > a small file, and there's no apparent resource crunch going on. I > > think I can rule this out. Is there any other reason the kernel > > itself might issue KILL signals? > > > > How can I get information on what's going on? Is there some way I > > can put in a trace on any KILL signals issued on the system so I > > can identify the culprit process (or kernel)? I can get a list of > > killed processes from the system accounting (lastcomm), but I need > > the signal type and source. > > > > Any ideas? > > Giving us the version of FreeBSD you're running usually helps, with > that noted there was a pretty serious bug a while back where some > system daemon would incorrectly send signals to processes it > shouldn't have. > > I would recommend an upgrade to -stable. > syslogd was the guilty party, killing of processes. I think it only happened when pids were recycled and you are using the '|command' form in syslog.conf. That bug gave me a lot of headaches back then, but is fixed in 3.3 Stable. Gerd To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message