Release criteria proposal: downgrade some kickstart delivery methods from Beta to Final

[Adam Williamson]

Hey, folks. So, currently the Beta criteria state:

"The installer must be able to use all kickstart delivery methods"

This is probably over-ambitious for Beta. We have some pretty odd
kickstart delivery methods:

https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg

that are only really useful in pretty unusual scenarios. In fact the
first of these is broken in F16 Beta and we decided to go ahead and
release it anyway (on the basis that we agreed this criterion should be
changed, which is why I'm proposing a change now), and the world has not
ended.

I'd propose at least this much change:

for Beta, the criterion should read

"The installer must be able to use the HTTP and NFS kickstart delivery
methods"

as those are the two that are really useful in most situations, and we
move

"The installer must be able to use all kickstart delivery methods"

to be a Final criterion. Thoughts? The Beta criterion is a bit more
'technology-specific' than I usually like to make the criteria, but I
don't think we're that likely to discover any exciting new protocols in
the foreseeable future, so specifying HTTP and NFS is probably
reasonably future-proof.

[Stephen John Smoogen]

On Tue, Oct 11, 2011 at 19:44, Adam Williamson  wrote:
> Hey, folks. So, currently the Beta criteria state:
>
> "The installer must be able to use all kickstart delivery methods"
>
> This is probably over-ambitious for Beta. We have some pretty odd
> kickstart delivery methods:
>
> https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
> https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg
>
> that are only really useful in pretty unusual scenarios. In fact the
> first of these is broken in F16 Beta and we decided to go ahead and
> release it anyway (on the basis that we agreed this criterion should be
> changed, which is why I'm proposing a change now), and the world has not
> ended.
>
> I'd propose at least this much change:
>
> for Beta, the criterion should read
>
> "The installer must be able to use the HTTP and NFS kickstart delivery
> methods"
>
> as those are the two that are really useful in most situations, and we
> move
>
> "The installer must be able to use all kickstart delivery methods"
>
> to be a Final criterion. Thoughts? The Beta criterion is a bit more
> 'technology-specific' than I usually like to make the criteria, but I

For those of us not up on all the methods these days.. what are they
(and which one didn't work for Beta :)) The reason I ask is if one of
the methods is via UUCP .. I don't think even Final is really
something we could care about :).

[Adam Williamson]

On Tue, 2011-10-11 at 20:05 -0600, Stephen John Smoogen wrote:
> On Tue, Oct 11, 2011 at 19:44, Adam Williamson  wrote:
> > Hey, folks. So, currently the Beta criteria state:
> >
> > "The installer must be able to use all kickstart delivery methods"
> >
> > This is probably over-ambitious for Beta. We have some pretty odd
> > kickstart delivery methods:
> >
> > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
> > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg
> >
> > that are only really useful in pretty unusual scenarios. In fact the
> > first of these is broken in F16 Beta and we decided to go ahead and
> > release it anyway (on the basis that we agreed this criterion should be
> > changed, which is why I'm proposing a change now), and the world has not
> > ended.
> >
> > I'd propose at least this much change:
> >
> > for Beta, the criterion should read
> >
> > "The installer must be able to use the HTTP and NFS kickstart delivery
> > methods"
> >
> > as those are the two that are really useful in most situations, and we
> > move
> >
> > "The installer must be able to use all kickstart delivery methods"
> >
> > to be a Final criterion. Thoughts? The Beta criterion is a bit more
> > 'technology-specific' than I usually like to make the criteria, but I
>
> For those of us not up on all the methods these days.. what are they
> (and which one didn't work for Beta :)) The reason I ask is if one of
> the methods is via UUCP .. I don't think even Final is really
> something we could care about :).

I linked the two weirder ones in my post. The more common ones are just
to retrieve the ks from an NFS server or HTTP server.

[Adam Williamson]

On Tue, 2011-10-11 at 18:44 -0700, Adam Williamson wrote:
> Hey, folks. So, currently the Beta criteria state:
>
> "The installer must be able to use all kickstart delivery methods"
>
> This is probably over-ambitious for Beta. We have some pretty odd
> kickstart delivery methods:
>
> https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
> https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg
>
> that are only really useful in pretty unusual scenarios. In fact the
> first of these is broken in F16 Beta and we decided to go ahead and
> release it anyway (on the basis that we agreed this criterion should be
> changed, which is why I'm proposing a change now), and the world has not
> ended.
>
> I'd propose at least this much change:
>
> for Beta, the criterion should read
>
> "The installer must be able to use the HTTP and NFS kickstart delivery
> methods"
>
> as those are the two that are really useful in most situations, and we
> move
>
> "The installer must be able to use all kickstart delivery methods"
>
> to be a Final criterion. Thoughts? The Beta criterion is a bit more
> 'technology-specific' than I usually like to make the criteria, but I
> don't think we're that likely to discover any exciting new protocols in
> the foreseeable future, so specifying HTTP and NFS is probably
> reasonably future-proof.

Ping on this one: any further thoughts? Didn't get much response the
first time out.

[Kamil Paral]

> On Tue, 2011-10-11 at 18:44 -0700, Adam Williamson wrote:
> > Hey, folks. So, currently the Beta criteria state:
> >
> > "The installer must be able to use all kickstart delivery methods"
> >
> > This is probably over-ambitious for Beta. We have some pretty odd
> > kickstart delivery methods:
> >
> > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
> > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg
> >
> > that are only really useful in pretty unusual scenarios. In fact
> > the
> > first of these is broken in F16 Beta and we decided to go ahead and
> > release it anyway (on the basis that we agreed this criterion
> > should be
> > changed, which is why I'm proposing a change now), and the world
> > has not
> > ended.
> >
> > I'd propose at least this much change:
> >
> > for Beta, the criterion should read
> >
> > "The installer must be able to use the HTTP and NFS kickstart
> > delivery
> > methods"
> >
> > as those are the two that are really useful in most situations, and
> > we
> > move
> >
> > "The installer must be able to use all kickstart delivery methods"
> >
> > to be a Final criterion. Thoughts? The Beta criterion is a bit more
> > 'technology-specific' than I usually like to make the criteria, but
> > I
> > don't think we're that likely to discover any exciting new
> > protocols in
> > the foreseeable future, so specifying HTTP and NFS is probably
> > reasonably future-proof.
>
> Ping on this one: any further thoughts? Didn't get much response the
> first time out.

I agree.

[John Dulaney]

> Date: Wed, 16 Nov 2011 08:44:52 -0500
> From: [e-mail]
> To: [e-mail]
> Subject: Re: Release criteria proposal: downgrade some kickstart delivery methods from Beta to Final
> > On Tue, 2011-10-11 at 18:44 -0700, Adam Williamson wrote:
> > > Hey, folks. So, currently the Beta criteria state:
> > >
> > > "The installer must be able to use all kickstart delivery methods"
> > >
> > > This is probably over-ambitious for Beta. We have some pretty odd
> > > kickstart delivery methods:
> > >
> > > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_File_Path_Ks_Cfg
> > > https://fedoraproject.org/wiki/QA:Testcase_Kickstart_Hd_Device_Path_Ks_Cfg
> > >
> > > that are only really useful in pretty unusual scenarios. In fact
> > > the
> > > first of these is broken in F16 Beta and we decided to go ahead and
> > > release it anyway (on the basis that we agreed this criterion
> > > should be
> > > changed, which is why I'm proposing a change now), and the world
> > > has not
> > > ended.
> > >
> > > I'd propose at least this much change:
> > >
> > > for Beta, the criterion should read
> > >
> > > "The installer must be able to use the HTTP and NFS kickstart
> > > delivery
> > > methods"
> > >
> > > as those are the two that are really useful in most situations, and
> > > we
> > > move
> > >
> > > "The installer must be able to use all kickstart delivery methods"
> > >
> > > to be a Final criterion. Thoughts? The Beta criterion is a bit more
> > > 'technology-specific' than I usually like to make the criteria, but
> > > I
> > > don't think we're that likely to discover any exciting new
> > > protocols in
> > > the foreseeable future, so specifying HTTP and NFS is probably
> > > reasonably future-proof.
> >
> > Ping on this one: any further thoughts? Didn't get much response the
> > first time out.
>
> I agree.

I'm also +1 on this.

[Jared K. Smith]

On Tue, Nov 15, 2011 at 10:07 PM, Adam Williamson  wrote:
> Ping on this one: any further thoughts? Didn't get much response the
> first time out.

+1 from me.

--
Jared Smith
Fedora Project Leader