Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!pasteur!ucbvax!BU-CS.BU.EDU!bzs
From: bzs@BU-CS.BU.EDU (Barry Shein)
Newsgroups: comp.sys.workstations
Subject: Re: Configuring >10 Sun clients
Message-ID: <19763@bu-cs.BU.EDU>
Date: 7 Feb 88 23:03:49 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 35
Approved: works@rutgers.edu

From: wbkh@cgch.UUCP (Bruce K. Haddon)
>The Sun Microsystems Configuration Guide states that a "range of 1 to
>20 client workstations per disk server" are permitted, and that with
>"one 575-megabyte disk, 1 to 10 clients" are possible.
>
>Our situation is that we have tried with a Sun-3/280, with two
>SuperEagle 575-Mbytes disks (with a controller each), to configure
>more that 10 clients.  But if we specify more that 10 file systems in
>/etc/fstab, the system objects (I do not have the actual messages and
>symptoms to hand).  We really do need to go beyond the 10 limit, and
>with the amount of disk that we have, there should be no trouble
>going as far as 20.

Whoa, slow down, park it right over there and let's talk...

What does "file systems in fstab" have to do with setting up diskless
clients? They don't use separate file systems in fstab.

Their swap and root are specified in /etc/nd.local as ranges of blocks
which split up a very few (usually one per physical disk on the
server, two Super Eagles, two partitions) partitions. You can have
more but it's not particularly helpful or necessary (usually the
result of an unplanned expansion.)

The Sun setup utility should lead you through this although you will
probably have to do a full dump of your entire system and start over
again unless you really know how to set these things up yourself
(doesn't sound like it, but you could learn, heh heh.) I recently
posted a note in SUN-SPOTS (comp.sys.sun) describing the manual
procedure in general terms.

In short: the partitions for the diskless clients are not reflected in
/etc/fstab (other than /pub), they're in /etc/nd.local in an entirely
different form. Of course, any observed complexity is "fixed in 4.0".

	-Barry Shein, Boston University