Subject: Installing via build.sh.
To: None <current-users@netbsd.org>
From: Richard Rauch <rkr@olib.org>
List: current-users
Date: 05/22/2003 17:17:10
I hit an odd problem: I did "./build.sh distribution" and then also "sets",
on my fastest machine, then as root on the laptop, over an NFS mount, I
attempted to do a "./build.sh install".  This failed over a missing "pax"
("nbpax"?) binary.

Also, everytime I run build.sh, nbmake seems to get rebuilt.  This doesn't
take *too* long, but is a bit annoying.  (Yes, I guess I could just build
sets and manually unpack them onto the laptop...)

Is it supposed to work to "./build.sh distribution && ./build.sh install=/"?
Is there any reason why it should fail if the "install" is on another
machine?

For whatever it's worth, both machines are running NetBSD 1.6, differing
only in kernel configs and a tiny hack to the PCMCIA ne driver for the
laptop.

I guess for now I'll just untar everything but etc.tgz by hand on the
laptop, but I'd like to be able to do this using build.sh.


Thanks.


 -----

A quick network topology, for concreteness:

                         [anoncvs]
                             |
 odysseus  --- NFS --->    hermes
 (laptop)                 (athlon)
                             |
                          local fs
                             |
                  ~rkr/netbsd/current/src

(Technically, "odysseus" is "podysseus" on the "Private" ethernet,
while "hermes" is both "hermes" and "phermes"...(^&)

-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/