From owner-freebsd-chat Sun Jan 26 13:52:12 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id NAA26854 for chat-outgoing; Sun, 26 Jan 1997 13:52:12 -0800 (PST) Received: from agora.rdrop.com (root@agora.rdrop.com [199.2.210.241]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id NAA26777 for ; Sun, 26 Jan 1997 13:50:49 -0800 (PST) Received: from linkdead.paranoia.com by agora.rdrop.com with smtp (Smail3.1.29.1 #17) id m0vocTG-0008xZC; Sun, 26 Jan 97 13:50 PST Received: from localhost (localhost [127.0.0.1]) by linkdead.paranoia.com (8.7.6/8.7.3) with SMTP id PAA19513; Sun, 26 Jan 1997 15:38:13 -0600 (CST) Message-Id: <199701262138.PAA19513@linkdead.paranoia.com> X-Authentication-Warning: linkdead.paranoia.com: Host localhost [127.0.0.1] didn't use HELO protocol To: netbsd-users@netbsd.org, freebsd-chat@freebsd.org Subject: The wonderful world of heterogenous PC OSes Date: Sun, 26 Jan 1997 15:38:13 -0600 From: VaX#n8 Sender: owner-chat@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Nontechnical: Hi. While getting a second system up and running today, I was contemplating that I seem to be gathering one computer per OS. This seemed rather silly to me, since one PC and a cool boot manager aught to do the trick, but I've begin to find out a couple of things: Some of them boot in profoundly dumb ways (these are mostly commercial ones) Some of them don't coexist peacefully Unix likes to stay running; named caches names and a reboot blows the cache away (I have an idea of sending it SIGINT and saving the /var/tmp file away in some kind of cache for next reboot, would that work?) Shutting down frequently leaves jobs half-finished. There's often no good recovery. There's not even a particularly good way to (for example) have Unix change personalities and start PPPing into somewhere else, acting like a different host (sendmail caches the hostname and you run into other problems, and of course a firewalled setup gets complicated). People like to have a box running; that way you can debug your problem with the help of the Internet, and if you screw up the boot process or partitioning, your important stuff is still safe. Therefore, it seems logical to have a Unix "server" which has most of the storage and media types, changes infrequently both hardware-wise and system-sw-wise, and has a very simple setup (e.g. no need in having 4 OSes on it). Is this consistent with other people's collected wisdom? Anyone have comments or solutions I may have overlooked?