From owner-freebsd-sparc64@FreeBSD.ORG Sun Nov 4 22:45:32 2007 Return-Path: Delivered-To: freebsd-sparc64@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 7F9D716A421; Sun, 4 Nov 2007 22:45:32 +0000 (UTC) (envelope-from royce@alaska.net) Received: from malik.acsalaska.net (malik.acsalaska.net [209.112.173.227]) by mx1.freebsd.org (Postfix) with ESMTP id B357E13C48E; Sun, 4 Nov 2007 22:45:29 +0000 (UTC) (envelope-from royce@alaska.net) Received: from [192.168.254.100] (209-112-218-167-rb1.nwc.dsl.dynamic.acsalaska.net [209.112.218.167]) by malik.acsalaska.net (8.14.1/8.14.1) with ESMTP id lA4MjCOI011769; Sun, 4 Nov 2007 13:45:12 -0900 (AKST) (envelope-from royce@alaska.net) Message-ID: <472E4B8D.2020902@alaska.net> Date: Sun, 04 Nov 2007 13:45:33 -0900 From: Royce Williams User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.6) Gecko/20070728 Thunderbird/2.0.0.6 Mnenhy/0.7.5.0 MIME-Version: 1.0 To: Kris Kennaway References: <200711011822.25884.linimon@lonesome.com> <472B39A8.3070708@alaska.net> <472DAFBE.9070603@FreeBSD.org> In-Reply-To: <472DAFBE.9070603@FreeBSD.org> X-Enigmail-Version: 0.95.5 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-ACS-Spam-Status: no X-ACS-Scanned-By: MD 2.63; SA 3.2.3; spamdefang 1.122 Cc: Mark Linimon , freebsd-sparc64@freebsd.org Subject: hardware and package builds (was: Re: Doesn't anything work around here?) X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 04 Nov 2007 22:45:32 -0000 Kris Kennaway wrote, on 11/4/2007 2:40 AM: > That is no problem, even on single CPU machines I run concurrent builds > since it turns out to be more efficient. In the past we have even used > 14 CPU e4500 machines for package builds although they all died from > hardware failure. I'd like to put some time and money where my mouth is on package builds. To summarize a couple of threads and off-list exchanges: - A 4x 450MHz Ultra 80 would be good for sparc64 package builds, and all of the CPUs would be useful. - There are more machines used by the project than there is space to comfortably host them, so remote hosting is needed. - If more packages were working and built more often, and if freebsd-update was available for sparc64, the overall "health" of the platform would be improved. Based on the above, here are some proposed actions and associated questions. System: I am willing to buy a 4x U80 for myself and make it available to the project for package builds. Shipping to Alaska is a bear, so once the deal goes down, if any other donors could chip in to get it shipped here, I would appreciate it. I don't want to do this, however, unless it will actually be useful -- and used. Will it? Alternate system: IIRC from a couple of years ago, the unsupported onboard SCSI controller in my Ex500 machines prevented me from easily making them available to the project. What is the status of the needed (esp?) driver port from NetBSD? Disk: How much disk will be needed for package builds, and what configuration will reduce turnaround time? Remote access: I will ask my employer if we can host the U80 on work premises and provide remote access. Kris, I assume that your needs haven't changed since this post? http://lists.freebsd.org/mailman/htdig/freebsd-sparc64/2004-August/001961.html Load/use: How much daily/hourly load would the system see? I assume that ports are built on a rolling basis (as they're updated) between release cycles, and that there's a big push to prepare an entire package set for a given release? Tools: I've only just now discovered and started to try to grok ports/Tools/portbuild/scripts/*, so I have some catching up to do to even understand scope here. What other recommended reading is there? Royce -- Royce D. Williams - IP Engineering, ACS http://www.tycho.org/royce/ - PGP: 3FC087DB/1776A531 You can never tell what a programmer is doing 'til it's too late ~Cray