Subject: Re: binary snapshots?
To: <>
From: Michael L. VanLoon -- HeadCandy.com <michaelv@HeadCandy.com>
List: current-users
Date: 08/21/1995 23:41:08
Delete if you're not into nostalgia...

Earlier, I wrote:

>I figured I would look at upgrading my 1.0 system to current, after
>getting a new hard drive, and went about grabbing files to do so this
[...]
>I really don't think it's worth my time to try and upgrade through a
>source-only process (which I know can be quite messy).  I'd prefer to
>just do the "binary install, build kernel, build sources, you're done"
>approach.

And, in a self-serving 5 minutes of fame...  I was thinking back.  I
believe I know how messy this can be because: I believe *I* was the
first person to ever distribute a binary snapshot release of
NetBSD-current. :-) Thank you thank you.

It would have been in October or November of '92.  I had been itching
to update my 0.9 system to current, so I could get shared libraries.
Here, on my system, is the second shared library I built from
NetBSD-current (first being 1.0, of course; and, I had been running
NetBSD-0.8 since practically the day it came out, and 386BSD-1.0 and
0.0 before that -- this was just my first shot at running current):

    -r--r--r--  1 bin  bin  389354 Nov 19  1993 /usr/lib/libc.so.1.1

At the time, shared libraries was just the coolest thing in the whole
universe!  My office-mate and I both just couldn't get enough of
playing with them.  Especially since we had to use Ultricks for "real
work". :-)

I wanted to share my binaries for others because I wanted friends to
upgrade to current, and it was just the easiest way.  Of course, at
that time, NetBSD-current was a concept known only by a small, but
growing, group.  And, all sites were source-only archives for current.
It was expected (and not all that hard at that point) that you would
move from 0.9 to current by getting the sources and rebuilding parts
of your systems.

But it became harder over time, and I wanted to sell NetBSD to the
unwashed masses.  And, binary upgrades were so much easier.  So, I had
my poor little ALR BusinessVeisa, stingray.cc.iastate.edu, a 386/33,
w/5MB of RAM, two 100MB IDE hard drives, and a couple gig of space NFS
mounted from DECstations in my office all set up to go.  I was working
as a Systems programmer/admin type at Iowa State at the time.  My
office mate (Michael Graff <explorer@iastate.edu>) and I took up
ownership of the NetBSD archives on ftp.iastate.edu (he later took on
the task of managing all aspects of Iowa State's ftp server).

I set up a script that would sup current early in the morning and
rebuild the sources each day on my little 386.  I had to do
incremental builds, of course, because it would take three days to do
a clean build on the little machine, with mostly NFS-mounted storage.
Then, I had a script on Iowa State's ftp server rsh over and suck
tar-balls of my entire system onto ftp.iastate.edu each evening.  The
binary snapshots that appeared there were literally live images of my
running system.  Thus was created the first NetBSD-current binary
snapshot.

Of course, we couldn't stop there.  At the same time, we became a
full-service NetBSD-0.8, 0.9 and current source mirror.  A couple
other current binary snapshot sites sprung up a few months after that.
Probably the most notable one (in the U.S.) being at umich.edu.

This is probably also the primary reason ftp.iastate.edu became one of
the busiest, if not the busiest, NetBSD ftp site on the Internet.

Anyway, I just thought it was cool to reminisce for a few minutes...
Maybe explorer will add his own insight to this, if you're not bored
to tears already.  Anyway, back to your usual drivel... :-)

-----------------------------------------------------------------------------
  Michael L. VanLoon                                 michaelv@HeadCandy.com
       --<  Free your mind and your machine -- NetBSD free un*x  >--
     NetBSD working ports: 386+PC, Mac, Amiga, HP300, Sun3, Sun4, PC532,
                           DEC pmax (MIPS R2k/3k), DEC/AXP (Alpha)
     NetBSD ports in progress: VAX and others...
-----------------------------------------------------------------------------