Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!samsung!think!snorkelwacker!bloom-beacon!eru!luth!sunic!mcsun!cernvax!chx400!forty2!mecazh!paul
From: paul@mecazh.UUCP (Paul Breslaw)
Newsgroups: comp.sys.hp
Subject: Re^2: Root permission on NFS
Message-ID: <377@node17.mecazh.UUCP>
Date: 19 Feb 90 14:58:57 GMT
References: <376@node17.mecazh.UUCP> <11590023@hpubvwa.HP.COM>
Organization: Mecasoft SA, Zurich, Switzerland
Lines: 38

My original question was how to set root permissions on NFS mounted
file systems.

johnb@hpubvwa.HP.COM (John Blommers) writes:

> This is not the brightest thing to do to your servers, you know,
> since any root user on any client can now accidentally do a dread
> rm * on some part of the server file system, buying you big
> headaches.

I asked it because we have the following situation:-

At present we have a 340 cluster server with 1.1Gb of disk.

We have graduated to more diskless nodes than can be handled by
one cluster server, so have bought a second (and soon a third).

There is no need to split up our existing file systems. We are
adding another 1Gb of disk on the second server, and will split the
disk load between servers. Then we can provide a single 'virtual' 
file system, by cross-mounting the file systems.

Now I don't want to know which cluster server exports which file systems
in order to perform some administrative task. I simply want to do it
from whichever machine I happen to be logged in on (as root).

We do not have "any root user on any client", our diskless users are not
super users.

So what should I do?


Paul Breslaw
-- 
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Paul Breslaw, Mecasoft SA,          |  telephone :  41 1 362 2040
Guggachstrasse 10, CH-8057 Zurich,  |  e-mail    :  mcsun!chx400!mecazh!paul
Switzerland.                        |               paul@mecazh.UUCP