[1158] in NetBSD-Development
Installation changes and plans
daemon@ATHENA.MIT.EDU (Greg Hudson)
Sat Dec 16 17:50:31 1995
Date: Sat, 16 Dec 1995 17:50:08 -0500
From: Greg Hudson <ghudson@MIT.EDU>
To: netbsd-dev@MIT.EDU
I have made and tested the following changes to the installation
source tree:
* The server side of the installation is now split up into
seven scripts, called "disksetup", "makefs", "unpack",
"config", "athena-inst", "subscribe", and "mail". This
makes it a little bit easier to finish an installation if
you have to do some part by hand.
The scripts use /tmp/vars to communicate values with each
other; the beginning of each script specifies what /tmp/vars
inputs and outputs it uses.
* inetd.conf and ttys use ed scripts rather than add scripts.
The ed script for inetd.conf disables the normal telnetd,
and the ed script for ttys puts the ttyv6 line after the
ttyv5 line instead of at the end.
* If you create a new partition during the installation, the
script doesn't bother to ask you to confirm that you want to
overwrite it.
* The installation script now allows the user to get Kerberos
tickets and subscribe to netbsd-announce.
* The default is not to create an mfs partition for /tmp; the
wording suggests that users may want to do so if they have a
lot of memory or swap space.
* root's path includes some more directories, and its MANPATH
includes /usr/athena/man.
I have not updated the actual installation; that will wait until the
following happens:
* I move the installation source out of the NetBSD source tree
and into /mit/netbsd/dev/install. This step will probably
happen tonight.
* I examine all of the changes we've made to the
netbsd-current source tree, and catalog and propagate the
ones that need to be applied to the netbsd-1.1 tree.
In particular, I need to resolve the issue that NetBSD
uses hard links between directories, which is not supported
by AFS. As long as we want to have full system packs in
AFS, we'll need to do a translation to symbolic links or
copies at some point.
* I make an "mit-install" (separate from "athena-install")
with the MIT networking changes, so that our actual tarballs
can be as vanilla as possible.
* I set aside a space (possibly in AFS, possibly in NFS,
possibly on local disk) for building the NetBSD tree to make
tarballs from. This may be the same as the new system packs
location, or it may be different.
* I build a new /usr/athena, and possibly make a snapshot of
the sipb-athena source tree
* We do testing for a little while, and freeze the release and
source trees.
Hopefully, this can all happen before jhawk gives the IAP class,
possibly at the price of cutting testing short.
