From owner-freebsd-current Fri May 17 13:27:18 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id NAA16106 for current-outgoing; Fri, 17 May 1996 13:27:18 -0700 (PDT) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id NAA16067; Fri, 17 May 1996 13:27:03 -0700 (PDT) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id NAA20585; Fri, 17 May 1996 13:22:47 -0700 From: Terry Lambert Message-Id: <199605172022.NAA20585@phaeton.artisoft.com> Subject: Re: Re(2): Standard Shipping Containers - A Proposal for Distributing FreeBSD To: rkw@dataplex.net (Richard Wackerbarth) Date: Fri, 17 May 1996 13:22:46 -0700 (MST) Cc: freebsd-current@FreeBSD.org, hackers@FreeBSD.org, freebsd-stable@FreeBSD.org, nate@sri.MT.net In-Reply-To: from "Richard Wackerbarth" at May 16, 96 10:59:43 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > > Not true. If you have direct access to freefall (developers only), you can > use (4-sup) to get "up to the minute" copies of the CVS tree. > > If YOU can get "up to the minute" updates via sup, it is only because you fall > in my category (1). My proposal does not affect a sup server that does not > provide synchronous snapshots. If the mirror servers fired three times a day, and your pull from the mirror fired once a day, you'd be set. You only need an "up to the minute" version if that's how you locally maintain your changes -- by communicating them through the master SUP server. This assumes commit privs. If you don't have commit privs, once a day or once a week (depending on your level of activity) is enough to handle it for you. You can cut the intermediate sup server traffic by 33% if you institute multiple reader/single writer locks and use a pseudo writer lock for the SUP area mirroring. If commiters follow the "it must run before you release the writer lock" protocol, then you will be guaranteed a buildable image on every SUP. > > Since your assumptions are invalid for one of the two most common > > distribution method, the rest of the proposal is not completely valid. > > Since those who have the direct access are not really inhibited by this > proposal, I suggest that you reconsider it in view of the other 99.99% of the > folks for whom my assumptions apply. I don't totally agree with all aspects of the proposal, since I have multple trees for multiple concurrent projects, and it won't help me out that much because of it. But this is a point in its favor. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.