[Aurora-sparc-devel] Install Cd's?

[Dean Anderson]

On Thu, 10 Apr 2008, Tom "spot" Callaway wrote:

> On Wed, 2008-04-09 at 23:23 -0400, Dean Anderson wrote:
> > If you want to install a general machine and build the aurora from
> > nothing but source, and post the terminal log of the entire operation,
> > from install to distibution cd's I'd be more confident.
> 
> No offense, but I have better things to do with the next 6 months.

This shouldn't take 6 months. It should take a few minutes of your time,
(kickstart) and then some waiting, then testing the CDs and collecting
the logs at the end.  A few hours of your time.

This is just my point though. You _have_ probably spent 6 months making
one build work for one distribution by tweaking the build machines and
build software.  To repeat the process from scratch would take you
another 6 months. The risk for us is that your build machines are lost,
corrupted, hacked, etc, it will take us 6 months to figure it out.

> The document you referenced on how to spin an installable tree is
> pretty much obsolete now. Use pungi.

But pungi is software that needs some configuration; 'out of the box' it
doesn't get you aurora 2.99 Cds. You haven't published the configuration
information so that I can replicate your pungi environment.

By contrast, the build instructions (though I just actually looked at
that link--it is quite old) build the fedora sources and get you a CD
that can be installed.  These fairly worked in aurora 1.0.

> Pungi passes these flags to make the SPARC ISO images:
> 
> -G', '/boot/isofs.b', '-B', '...', '-s', '/boot/silo.conf',
> '-sparc-label', '"sparc"'
> 
> The old hack in mkinitrd doesn't exist anymore because... it isn't
> needed. Silo added an isofs.b specifically to make that obsolete.

The point of all this is that what you are doing is different from the
x86 cds made by fedora, so you have a deviation from the fedora
instructions.  A deviation which you haven't previously documented.

Since you have some deviation, you can have any solution. There is more
than one way to fix a problem.  Maybe you fixed the problem by using the
old mkisofs. We don't know how you fixed the problem. If you are hit by
a train, we don't know what you did---you haven't documented it.  It
would take us 6 months to figure it out.

Thus, my contention that aurora is currently unreproducible. I suppose
that should be qualified to say that Aurora is unreproducible without 6
months of work. But that is a long time to wait to fix a serious
security breach or other disasterous event.  Further, it is totally
avoidable with just a bit more transparancy by the aurora team.


> Any other accusations that you'd like to make? I was nowhere near
> Dallas on that day. ;)

Of course not.  That was /uhhgg...thump/ ;-}

		--Dean


-- 
Av8 Internet   Prepared to pay a premium for better service?
www.av8.net         faster, more reliable, better service
617 344 9000