From owner-cvs-all Mon Jan 25 03:06:16 1999 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id DAA28478 for cvs-all-outgoing; Mon, 25 Jan 1999 03:06:16 -0800 (PST) (envelope-from owner-cvs-all@FreeBSD.ORG) Received: from octopus.originative.co.uk (originat.demon.co.uk [158.152.220.9]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id DAA28461 for ; Mon, 25 Jan 1999 03:06:05 -0800 (PST) (envelope-from paul@originative.co.uk) From: paul@originative.co.uk Received: by OCTOPUS with Internet Mail Service (5.5.1960.3) id ; Mon, 25 Jan 1999 11:04:34 -0000 Message-ID: To: jkh@zippy.cdrom.com, jb@cimlogic.com.au Cc: peter@netplex.com.au, nsouch@teaser.fr, dillon@apollo.backplane.com, cvs-committers@FreeBSD.ORG Subject: RE: CVS totally broken on freefall Date: Mon, 25 Jan 1999 11:04:31 -0000 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.1960.3) Content-Type: text/plain Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk > -----Original Message----- > From: Jordan K. Hubbard [mailto:jkh@zippy.cdrom.com] > Sent: Sunday, January 24, 1999 11:20 PM > To: John Birrell > Cc: peter@netplex.com.au; nsouch@teaser.fr; > dillon@apollo.backplane.com; > cvs-committers@FreeBSD.org > Subject: Re: CVS totally broken on freefall > > > > This is a *VERY* good example of what not to do on a > production system. > > People should know better. Grrr. > > You know of a better way of upgrading machines? If so, tell me now > before we do hub. :) > > I don't have full spares for each, so I can't do the old roll-over > from a hot spare trick. If this were a commercial shop I'd have one > or two super-spares in the corner, sure, but project funds are not > quite so abundant. :) That's fair enough, this isn't a commercial setup and some corners get cut but there are still some things that definately should have been done. 1) Schedule the upgrade and warn people about it. 2) Get a checklist of what services run on the machine and make sure they're going to still work afterwards. 3) Disable critical services, like cvs, while the upgrade takes place to prevent any mishaps 4) Disable logins while you're actually doing it so people don't do things in the middle of the upgrade. 5) Take a backup before starting. 6) Actually announce that it's taking place so everyone doesn't run around trying to work out why things have stopped working. The above steps just take organisation and would have prevented the cvs problems and the perl script problems that occured. Who knows what else might still fall out in the wash. Paul. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message