From owner-freebsd-current  Wed Jul 31 18:30:40 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id SAA18580
          for current-outgoing; Wed, 31 Jul 1996 18:30:40 -0700 (PDT)
Received: from ice.fit.qut.edu.au (ice.fit.qut.edu.au [131.181.2.9])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id SAA18575;
          Wed, 31 Jul 1996 18:30:35 -0700 (PDT)
Received: (from tony@localhost) by ice.fit.qut.edu.au (8.7.5/8.7.3) id LAA21976; Thu, 1 Aug 1996 11:30:58 +1000 (EST)
Date: Thu, 1 Aug 1996 11:30:56 +1000 (EST)
From: Tony Jago <tony@fit.qut.edu.au>
To: freebsd-current@freebsd.org
Subject: NFS Diskless Dispare...
Message-ID: <Pine.HPP.3.92.960801105417.5061Q-100000@ice.fit.qut.edu.au>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-Sender: T.Jago@fit.qut.edu.au
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


 Hi there all, I have been attempting to run a diskless freebsd system in
 a student lab enviroment. The way the system is setup is that the client
 boots his PC with the netboot.com binary. This mounts a read-only root
 file system which does some basic configuration and then sends a message
 to the server which gets the server to setup a read-write file system for
 that client only. The client then mounts this new file system read-write
 and then does a chroot to this file system and keeps booting. /usr, /dev
 etc. are mounted read-only.

 This all works file when there are only a few machines booting but as the
 load increases problems start to occurr. I will list them below:

  1. The inability to mount file systems. The clients start barfing with
     something like "RPC mount timeout". This problem goes away after a
     while as the clients retry. I think its the mountd getting too many
     requests at once. Each client mounts 9 file systems.

  2. Files permissions are read incorrectly. Files that should be able to
     be executed are giving "permission denied" messages. Sometimes even
     the kernel can't be loaded by netboot.com but if you persist by
     typing "autoboot" it will magically start to work. Machines fail to
     boot correctly as programms called in /etc/rc don't start
     (permission denied).

  3. Pageing in of binaries cause the system to panic. Vnode_pager does
     not seem to like it when it can't page in executables, even when the
     file systems are hard mounted. Typically init can't be paged in and
     it Segment faults. This of course causes the kernel to panic. Other
     programs such as X often fall over but don't cause the system to
     panic. When this happens it happens to a number of machines in the
     room at exactly the same time. Perhaps this could be linked in with
     problem 2???

  The system is currently running split over 2 servers. I am running NFS
  version 2 with hard mounts. I tryed NFSv3 (both TCP and UDP) for a while
  but it was even worse. The servers have 16 nfsd's running and the
  clients have 4 nfsiod's. The most stable configuration I have found at
  this stage has been 2.2-960612-SNAP but I am open to suggestions.

  I would also like to hear any success/failure stories from others
  running FreeBSD diskless systems.

  Thanks for taking the time to read to the bottom of this message.

---
Tony Jago, System Administrator,    E-Mail: T.Jago@fit.qut.edu.au
Faculty of Information Technology,  Web: http://www.fit.qut.edu.au/staff/~tony
Queensland University of Technology.
Box 2434, Brisbane 4001, AUSTRALIA.              "We need more horsepower!"
Phone: +61 7 3864-2573 Fax: +61 7 3864-1959