Xref: utzoo comp.unix.admin:1582 news.admin:13405
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!think.com!mintaka!ogicse!unicorn!henson!jeff
From: jeff@henson.cc.wwu.edu (Jeff Wandling)
Newsgroups: comp.unix.admin,news.admin
Subject: Remotely mounted news. What are the strategies?
Summary: News is on server, how much of the FS tree should I mount on client?
Keywords: NFS news nn rn /etc/exports /etc/fstab nfssetup
Message-ID: <1991Apr12.175554.6409@henson.cc.wwu.edu>
Date: 12 Apr 91 17:55:54 GMT
Article-I.D.: henson.1991Apr12.175554.6409
Distribution: comp,news
Organization: Western Washington University
Lines: 40


Systems: DecStation 5500 (server) & DS 5000/200's (clients) Ultrix 4.1 (all)

   We have our news (nntp, nn, rn, etc..) on our server "henson". We also have 
several other machines, all clients of henson where *all* user files are
remotely mounted from henson to all the clients via NFS. Henson is where the
files actually "live". On clients, like "beaker", "rowlf", "fozzie", or
"gonzo", the user sees:

% mount
/dev/rz0a on / type ufs
/dev/rz0g on /usr type ufs
[... misc user file trees ...]
henson:/usr/local on /usr/local type nfs (ro,hard,bg)
henson:/usr/spool/mail on /usr/spool/mail type nfs (rw,hard,bg)
%

   In order to remotely mount the news (binaries and data) on the clients,
would it be wise to just mount the entire "/usr" tree, OR mount each of the
directories used by news, (/usr/lib/news, /usr/local/lib/nn, /usr/spool/news,
etc...) remotely from henson to each of the clients via NFS.

   What are the more popular (and wiseer) strategies to do this? 

   We have other programs line TeX, magic, etc.., that will need to 
be mounted. The more software subsets I see that I have to mount, 
the more I think `just mount the whole "/usr" subtree and not mess with 
"each little directory"'

   Is that strategy not taking full advantage of NFS? Is there some feature
of NFS that I don't recognize that would make this decision easier? or make
the process seem clearer?

   Any replies would be appreciated.



-- 
     Jeff Wandling        
     jeff@henson.cc.wwu.edu