From owner-freebsd-stable Sun Sep 1 11:46:54 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 CB73237B400 for ; Sun, 1 Sep 2002 11:46:49 -0700 (PDT) Received: from earth.hub.org (earth.hub.org [64.49.215.11]) by mx1.FreeBSD.org (Postfix) with ESMTP id 80C5043E6A for ; Sun, 1 Sep 2002 11:46:49 -0700 (PDT) (envelope-from scrappy@hub.org) Received: from earth.hub.org (earth.hub.org [64.49.215.11]) by earth.hub.org (Postfix) with ESMTP id E5BA92CC801; Sun, 1 Sep 2002 15:46:47 -0300 (ADT) Date: Sun, 1 Sep 2002 15:46:47 -0300 (ADT) From: "Marc G. Fournier" To: Matthew Dillon Cc: freebsd-stable@FreeBSD.ORG Subject: Re: Crash with KVM monitoring in place ... In-Reply-To: <200209011829.g81ITRWK014499@apollo.backplane.com> Message-ID: <20020901153933.S27764-100000@hub.org> 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 On Sun, 1 Sep 2002, Matthew Dillon wrote: > :Okay, according to my logs, here is what I have: > : > :Aug 31 22:13:47 venus scrappy[35952]: Starting KVM Monitoring > :Aug 31 22:13:47 venus scrappy[35962]: kernel_vm_end: 0xf6400000 > :Aug 31 23:00:03 venus scrappy[41669]: kernel_vm_end: 0xf6800000 > :Sep 1 00:00:25 venus scrappy[50135]: kernel_vm_end: 0xf7000000 > : > :Note that I modified my script so that it only reports to syslog when the > :value changes, instead of every minute ... > : > :On the other server, she's still running with: > : > :Sep 1 02:05:30 jupiter root[93918]: kernel_vm_end: 0xfa000000 > :Sep 1 02:07:32 jupiter root[95929]: kernel_vm_end: 0xfa400000 > :Sep 1 02:09:34 jupiter root[98699]: kernel_vm_end: 0xfa800000 > :Sep 1 02:10:34 jupiter root[1076]: kernel_vm_end: 0xfac00000 > :Sep 1 04:30:42 jupiter root[33295]: kernel_vm_end: 0xfb000000 > : > :Both server(s) are running kernels updated to sources of last night. > : > :Next? > > A kernel core is what is really needed. All I can suggest > without a kernel core is to perhaps turn on DDB and instead > of it crashing and rebooting you drop into DDB and do a > 'trace'. If over several crashes it dies in the same place > then we at least have an idea where to look. > > How large is your swap space, or your largest free partition? My swap is only ~2gig, and my drive looks like: venus# df -t ufs Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/amrd0s1a 103777974 68011064 27464674 71% / The server is remote, which makes most "normal debugging" quite difficult to do :( Awhile back, someone posted this to the lists: http://www.cs.duke.edu/~anderson/freebsd/netdump/readme.html To do dumps to a 'dump server' ... i tried it on a 4.4 kernel, I believe it was, and altho it looked like it was dumping, I got zero bits on the remote server, so I'm figuring changes in the kernel itself ... Any chance on getting such working (and maybe into) the latest source trees? It seems to be it would be most valuable, especially as more and more "large servers" are coming online? I've got the two servers sitting next to each other, so network isn't an issue, and they are both running fxp interfaces ... the code itself is at: http://www.cs.duke.edu/~anderson/freebsd/netdump/ with the netdump_client.c code being 17k, and the _server.c code being 7k, so they aren't that large .. the code appears to be from '00, so don't know what would have to be changed to reflect -STABLE's present state ... To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message