Managing a network of UNIX workstations

[Steve Simmons]

barrett at jhunix.HCF.JHU.EDU (Dan Barrett) writes:

>	I may be managing a network of DECstation 3100's running Ultrix in
>the near future.  I have been managing VAXen for a long time, but never a
>network of workstations.  So, I have some questions:

[[and goes on to ask a host of questions indicating he's really thought
  about the problems.]]

As a pre-emptive strike, you should immediately order a copy of the
back proceedings of the USENIX Large Installation and Systems Admin
conference proceedings.  There have been three, order them from the
Usenix Association, 2650 Ninth St, Suite 215, Berkeley, CA, 94710.
Lisa I is $4.00, Lisa II is $8.00, Lisa III is $13.00.  I and II are
photocopy only, III is bound.  I believe there will be a Lisa IV
in Monteray in 1990, but I've seen no formal announcement.  The LISA
proceedings aren't perfect by any means, but you can get a wide
variety of ideas and contacts.

Also of good use is the newly-formed and still incomplete UNIX admin
archives.  It's still under construction, but if you can anon ftp
to terminator.cc.umich.edu, look under ~ftp/unix/sysadm.  We will
be keeping lots of odds and ends relating to sysadm stuff there,
see the readme file for details.  Anon uucp will be allowed fairly
soon.  Full details when it's all set up -- maybe another week.

>(2)	How do you do transparent backups?  I want to pop a tape in ONE
>	tape drive and say "Back up ALL files from ALL workstations onto
>	this tape."

Aside from security issues, where the heck are you going to get a
tape that big?

>(3)	We'd like all users to have accounts on all workstations.  What's
>	the best way to maintain an inter-machine password file?  I've
>	heard vaguely of "yellow pages" but have never used it.

YP will work but isn't real wonderful.  You might also look into
Hesiod from Project Athena.

>(4)	We'd like a system where the entire network appears to each user as
>	if it were one huge "machine".  A user would log onto this "machine"
>	and not care which workstation s/he were actually using.  (Maybe the
>	"machine" would automatically log the user onto the workstation with
>	the lightest system load.  I've seen this done with VMS systems at
>	other schools.)  Can this entire scheme be done?  Transparently?

Yes, it can be done -- I did it with an all-Sun network at Schlumberger
(170 modes) and will have it done in another month or so at ITI.  I will
use YP for login IDs, BIND for all other services.  The file systems will
be set up such that all users homes are in '/home/<hostname>[1-n]/user'.
This gives a consistant network view from everywhere.  Since you're all
3100s, architectural and operating system differences aren't a problem.

>(5)	Should we put disks on every workstation, or have one fileserver and
>	many diskless workstations?  Which is better?  Easier to maintain?

I like 'dataless' nodes -- enough disk so that the system can boot
itself and swap (the / partition + swap, basicly).  This cuts 90% of
the writes the file servers must do, optimising their performance.
Unfortunately DEC doesn't yet support booting from the 100MB internals.
So get swap disks, and when DEC supports booting from them, convert to
dataless.