From owner-freebsd-questions Thu Dec 3 18:34:58 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id SAA06263 for freebsd-questions-outgoing; Thu, 3 Dec 1998 18:34:58 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from sturm.canonware.com (canonware.com [204.107.140.54]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA06256 for ; Thu, 3 Dec 1998 18:34:56 -0800 (PST) (envelope-from jasone@canonware.com) Received: from canonware.com (mg-20425418-99.ricochet.net [204.254.18.99]) by sturm.canonware.com (8.8.8/8.8.8) with ESMTP id SAA29840; Thu, 3 Dec 1998 18:30:21 -0800 (PST) (envelope-from jasone@canonware.com) Message-ID: <36674A29.21201E33@canonware.com> Date: Thu, 03 Dec 1998 18:34:17 -0800 From: Jason Evans X-Mailer: Mozilla 4.07 [en] (X11; I; FreeBSD 2.2.7-STABLE i386) MIME-Version: 1.0 To: Greg Lehey CC: freebsd-questions@FreeBSD.ORG Subject: Re: crashdump, dangerously dedicated, hosed system References: <19981204125003.M441@freebie.lemis.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Greg Lehey wrote: > There's nothing obvious you did wrong, though there's a possibility > that something may have gone wrong if your swap partition was exactly > the size of your memory (offhand I can't remember, but it needs a > little bit more to at least note the fact that there's a dump in the > partition). Also, of course, if you had laid out your drive with > overlapping partitions, all bets would be off. Finally, depending on > how your system crashed, that could have caused the corruption. I > don't suppose you'll be able to find out now. It's very rare to have > this kind of problem; I've been running BSD on multiple machines for > nearly 7 years, and it's never happened to me. It appears to have been my fault. I can't say for sure, since I lost the disklabel, but it appears that my swap partition was too small. When I rebuilt the disk, I was careful to make it large enough, and had no problems capturing the crashdump thereafter. I didn't think about this being the potential problem because at the time I wasn't aware that the crashdump is dumped starting at the end of the swap partition and working down. (Lucky I did a backup just before forcing the crash. =) ) As for the cause of the crash, it's a bit embarrassing to admit. I was running out of mbuf clusters, even though I had maxusers set to 100 (2112 mbufs). It turns out my test code was sending two orders of magnitude more data than I had remembered, and in a manner that was pathologically bad as far as mbuf usage is concerned. Thanks for the email, Jason -- Jason Evans Email: [jasone@canonware.com] Web: [http://www.canonware.com/~jasone] Home phone: [(650) 856-8204] Work phone: [(415) 808-8742] Quote: ["Invention is 1% inspiration and 99% perspiration" - Thomas Edison] To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message