Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!mit-eddie!apollo!rees
From: rees@apollo.uucp (Jim Rees)
Newsgroups: news.admin
Subject: Re: NFS news - problem ?? -- really RFS news sharing
Message-ID: <379aea64.b8ab@apollo.uucp>
Date: Thu, 1-Oct-87 07:40:00 EDT
Article-I.D.: apollo.379aea64.b8ab
Posted: Thu Oct  1 07:40:00 1987
Date-Received: Mon, 5-Oct-87 02:17:37 EDT
References: <271@flmis06.ATT.COM> <539@nud.UUCP>
Organization: Apollo Computer, Chelmsford, Mass.
Lines: 17


    We use the pseudo machine name "ME" in the sys file and advertise NEWSLIB
    along with the database.  In fact, we put all the news related files into
    a single directory hierarchy (/news) and then advertize the single resource.
    As long as only one machine runs sendbatch, users on all machines can read
    and post news without problems.  We now have a dozen machines here sharing
    news through RFS.

We also use a single /news hierarchy for all our machines.  There are
some locking and contention problems that start to show up when you
reach several thousand nodes sharing the same /news (we have about 2200),
but these are easily fixed (I'm sending patches to Rick Adams).

To further simplify matters, our entire network appears as a single
uucp host ("apollo") to the outside world.  So the sys file (and most
everything else) is set up just the way it would be on a single
machine.