From owner-freebsd-stable@FreeBSD.ORG Thu Jun 5 05:19:06 2008 Return-Path: Delivered-To: freebsd-stable@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id AEA41106566C; Thu, 5 Jun 2008 05:19:06 +0000 (UTC) (envelope-from jrhett@netconsonance.com) Received: from mail.netconsonance.com (mail.netconsonance.com [198.207.204.4]) by mx1.freebsd.org (Postfix) with ESMTP id 845158FC18; Thu, 5 Jun 2008 05:19:06 +0000 (UTC) (envelope-from jrhett@netconsonance.com) Received: from [172.16.12.8] (covad-jrhett.meer.net [209.157.140.144]) (authenticated bits=0) by mail.netconsonance.com (8.14.1/8.14.1) with ESMTP id m555J34e079385; Wed, 4 Jun 2008 22:19:03 -0700 (PDT) (envelope-from jrhett@netconsonance.com) X-Virus-Scanned: amavisd-new at netconsonance.com X-Spam-Flag: NO X-Spam-Score: -3.691 X-Spam-Level: X-Spam-Status: No, score=-3.691 tagged_above=-999 required=3.5 tests=[ALL_TRUSTED=-1.44, AWL=-2.251] Message-Id: <13A3FC54-B459-48C5-85CD-14CC38913838@netconsonance.com> From: Jo Rhett To: Edwin Groothuis In-Reply-To: <20080605003545.GP89632@k7.mavetju> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v919.2) Date: Wed, 4 Jun 2008 22:19:03 -0700 References: <458FE12C-AE4D-48F9-8193-4663079CEEF8@netconsonance.com> <84EBEA5D3A1F47E79E8E12C4CF4D0314@multiplay.co.uk> <20080605003545.GP89632@k7.mavetju> X-Mailer: Apple Mail (2.919.2) Cc: FreeBSD Stable , Doug Barton , Steven Hartland Subject: Re: challenge: end of life for 6.2 is premature with buggy 6.3 X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 05 Jun 2008 05:19:06 -0000 On Jun 4, 2008, at 5:35 PM, Edwin Groothuis wrote: > Use the eat-your-own-food approach (while not knowing what the 500 > systems do): Make sure you use the same hardware and software as > what is in production. Upgrade it first, run it for two weeks. If > it doesn't, fallback and see where it went wrong. If it all works > fine after two weeks, roll it out. Edwin, I've been building testbed environments for over 20 years in my professional career. I know a lot more than this basic concept. The costs in our environment for a proper testbed is $20k in hardware and 3000 man hours. That's for a small test of comparable small changes to the existing environment. Why would we take on this cost only to re-document well known and already acknowledged bugs? I mean, really? Not trying to be sarcastic, but do you purchase cars to test them out and see if you can get better gas mileage than the EPA observes? Neither do I ;-) (yes, their testing methodology is flawed but it's a decent enough benchmark to know if you want the vehicle or not) -- Jo Rhett Net Consonance : consonant endings by net philanthropy, open source and other randomness