From owner-freebsd-questions Fri Sep 29 21:51:41 1995 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id VAA16749 for questions-outgoing; Fri, 29 Sep 1995 21:51:41 -0700 Received: from Root.COM (implode.Root.COM [198.145.90.17]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id VAA16743 for ; Fri, 29 Sep 1995 21:51:38 -0700 Received: from corbin.Root.COM (corbin [198.145.90.50]) by Root.COM (8.6.12/8.6.5) with ESMTP id VAA01262; Fri, 29 Sep 1995 21:50:16 -0700 Received: from localhost (localhost [127.0.0.1]) by corbin.Root.COM (8.6.12/8.6.5) with SMTP id VAA00295; Fri, 29 Sep 1995 21:51:35 -0700 Message-Id: <199509300451.VAA00295@corbin.Root.COM> To: Ron Lenk cc: freebsd-questions@freebsd.org Subject: Re: Mysterious reboot...any ideas? In-reply-to: Your message of "Fri, 29 Sep 95 18:14:53 MDT." <199509300014.SAA00425@widget.xmission.com> From: David Greenman Reply-To: davidg@Root.COM Date: Fri, 29 Sep 1995 21:51:34 -0700 Sender: owner-questions@freebsd.org Precedence: bulk >After doing a build of FreeBSD stable on Sept. 12th, my machine ran >smoothly for 13 days, before it rebooted mysteriously this afternoon, >while I was editing a message in emacs. > >At the bottom of the screen, I got a "Fatal error (1)" message, which >I later discovered came from emacs. Several seconds later ( not more >than 5 ) the familiar messages "...syslogd: exiting on signal 15", >followed by "syncing disks..." and "Rebooting..." were printed on >the console as if I had done a "reboot", or killed init. > >I was the only person logged in at the time, and I am the only one with >root access here, so I'm certain that nobody rebooted the machine on >purpose. It looks to me as if init just died. I've seen this once myself. It happend on wcarchive. It was weird - basically, somehow init either got sent a SIGTERM or it got confused. The result was that init killed everything off and went to single user mode. To this day I don't understand how this happend, but it was quite shocking when it occurred. All I did to restart things was type "umount -a; ^D", and the system came back up. Weird. I spent some time looking at the init sources and didn't see a problem there (but one could easily exist that I missed). My suspicion at the moment is that some system daemon has a bug (like an argument botch to kill(2)) that results in init getting sent a SIGTERM or other deadly signal. It might be useful if the kernel logged something via syslog whenever it sends a signal to init to indicate where the signal came from. -DG