Subject: Re: How do I build the boot fs?
To: Bruce Lane <kyrrin@bluefeathertech.com>
From: NetBSD Bob <nbsdbob@weedcon1.cropsci.ncsu.edu>
List: port-vax
Date: 12/17/2000 11:11:20
> 	Ok... now I'm -REALLY- confused. Why, if this is the case, did 1.4.1 and
> 1.4.2 have tape-bootable filesystems? (boot.fs.gz) Furthermore, why then do
> the installation instructions for NetBSD (including 1.5) explicitly
> reference a tape installation procedure?

The following is ONLY to raise issues for discussion and consideration,
NOT to come down on anyone in any way......

The new 1.4.3 release has the boot.fs too.

What has happened to the system code for VAXen between 1.4.x and 1.5?
A boot.fs or tape boot of some sort needs to be there.  Netbooting
is not always the ultimate solution.  I keep spare drives around to
boot from, now that I have VAXen up, but for folks just jumping in,
with a cold machine, it can be problematic.

1) If there is not a 1.5 boot.fs or tapeboot of some sort possible, we
   are shooting ourselves in the foote, and relegating old VAX users
   with MVII and older things to being second class citizens.
   IMHO, that is not good.

2) If there is not a 1.5 boot.fs or tapeboot of some sort possible, it
   is the end of the line developmentally for tape-drive VAXen.
   IMHO, that is not good.

3) The 1.5 system (great on fast VAXen) is abysmally slow in all
   respects on older things like MVII's.  I have NetBSD-1.2,
   NetBSD-1.4.3, NetBSD-1.5, and Ultrix up on my MVII for some
   general comparisons.  Ultrix is SLOW....granted, but it works.
   NetBSD-1.2 is light and zippy, and about 200% faster than Ultrix.
   NetBSD-1.4.3 is relatively light and equally zippy and about 220%
   faster than Ultrix.  NetBSD-1.5 is a behemoth, takes 500% longer
   to load, and only runs 150% faster than Ultrix.  These are only
   approximations comparing the systems on the same machine across
   identical HD's, but illustrate the point.  That point is that
   it seems we have, due to whatever bloat or other reasoning,
   overburdened the older VAXen relative to the 3100 and up newer
   VAXen.  Is there any solution to the problem?  If not, then
   NetBSD-1.4.3 seems to be the end of the line for older VAXen.

I know.... I should get a newer VAX (I do have several M38/M76 crates
though, but no 4000 and up boxes).  But the point is that we have begun
to diverge between the classes (oldeVAXen vs neweVAXen [the havenots
vs the haves?]).

I dunno what the answer is, but I sure feel the constraints of the
development into 1.5 as it now stands.

Any thoughts and discussion is appreciated.  Like I said, I am not
putting anyone down, but just raising issues that are beginning to
crop up, which may require solutions.

Thanks

Bob