MIA spins

[Paul W. Frields]

Despite a posting to the lists earlier in the month[1] to check our
spin content, apparently no one caught the fact that there was no
Moblin spin among the release candidates.  At this point, we're going
to have to wait for mirrors to die down before we can consider
shipping a late copy.

I went to the Spins process page[2] to see where things might have gone
off track, and found that the page stated that after acceptance the
following step occurs:

"Spin added to a wiki page containing a table listing all of the spins
for Fedora X (very similar to
https://fedoraproject.org/wiki/Releases/11/FeatureList)"

Note that no one is responsible or accountable for it happening
according to the spin page, nor is the proper page listed.  I made
a change, so now the step reads:

"The Spins Wrangler adds the Spin to a wiki page containing a table
listing all of the spins for Fedora 14
(https://fedoraproject.org/wiki/Releases/14/Spins)"

This change is not a decree by any means, but it seemed like the most
effective way to ensure entries get made.

I can't help but think we would have caught this problem if there were
regular Spins meetings, or at least schedule items that indicate the
SIG should carry out a task like checking the list of Spins before RC
phase.  Earlier, I brought up the idea of entering a set of schedule
items.  Ideally, someone on the SIG will step up to gather that list
of dates for Fedora 14, and talk to John Poelstra to make sure the
dates get on the release schedule.  I'd also like to make sure we have
a clear spins wrangler who knows the responsibilities that come with
that role.  That person should check over the spins process page[2]
and find any places where it's not clear what the task is, and who is
responsible for doing it.

So -- who wants to gather the schedule and report it to John Poelstra
by around June 1?

Who is the current spins wrangler, and can that person go over the
page by around the same time, and bring questions to the list?

* * *
[1] http://lists.fedoraproject.org/pipermail/spins/2010-May/001131.html
[2] https://fedoraproject.org/wiki/Spins_Process

-- 
Paul W. Frields                                http://paul.frields.org/
  gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233  5906 ACDB C937 BD11 3717
  http://redhat.com/   -  -  -  -   http://pfrields.fedorapeople.org/
          Where open source multiplies: http://opensource.com

[Peter Robinson]

Hi Paul,

On Wed, May 26, 2010 at 1:40 AM, Paul W. Frields <stickster(a)gmail.com&gt; wrote:
>  Despite a posting to the lists earlier in the month[1] to check our
>  spin content, apparently no one caught the fact that there was no
>  Moblin spin among the release candidates.  At this point, we're going
>  to have to wait for mirrors to die down before we can consider
>  shipping a late copy. 
What's the latest update of this? If the mirrors are still too busy
can we at least get the torrents up and propagating out with a link on
the site to them?

Peter

[Jesse Keating]

On Wed, 2010-05-26 at 16:02 +0100, Peter Robinson wrote:
>  Hi Paul,
>  
>  On Wed, May 26, 2010 at 1:40 AM, Paul W. Frields <stickster(a)gmail.com&gt; wrote:
>  > Despite a posting to the lists earlier in the month[1] to check our
>  > spin content, apparently no one caught the fact that there was no
>  > Moblin spin among the release candidates.  At this point, we're going
>  > to have to wait for mirrors to die down before we can consider
>  > shipping a late copy.
>  
>  What's the latest update of this? If the mirrors are still too busy
>  can we at least get the torrents up and propagating out with a link on
>  the site to them?
>   
I'm working on creating the spin isos today.

-- 
Jesse Keating
Fedora -- Freedom² is a feature!
identi.ca: http://identi.ca/jkeating

[Paul W. Frields]

On Wed, May 26, 2010 at 08:41:31AM -0700, Jesse Keating wrote:
>  On Wed, 2010-05-26 at 16:02 +0100, Peter Robinson wrote:
>  > Hi Paul,
>  > 
>  > On Wed, May 26, 2010 at 1:40 AM, Paul W. Frields <stickster(a)gmail.com&gt; wrote:
>  > > Despite a posting to the lists earlier in the month[1] to check our
>  > > spin content, apparently no one caught the fact that there was no
>  > > Moblin spin among the release candidates.  At this point, we're going
>  > > to have to wait for mirrors to die down before we can consider
>  > > shipping a late copy.
>  > 
>  > What's the latest update of this? If the mirrors are still too busy
>  > can we at least get the torrents up and propagating out with a link on
>  > the site to them?
>  > 
>  
>  I'm working on creating the spin isos today. 
Thanks, Jesse.

-- 
Paul W. Frields                                http://paul.frields.org/
  gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233  5906 ACDB C937 BD11 3717
  http://redhat.com/   -  -  -  -   http://pfrields.fedorapeople.org/
          Where open source multiplies: http://opensource.com

[Peter Robinson]

On Wed, May 26, 2010 at 9:41 PM, Paul W. Frields <stickster(a)gmail.com&gt; wrote:
>  On Wed, May 26, 2010 at 08:41:31AM -0700, Jesse Keating wrote:
> > On Wed, 2010-05-26 at 16:02 +0100, Peter Robinson wrote:
> > > Hi Paul,
> > >
> > > On Wed, May 26, 2010 at 1:40 AM, Paul W. Frields <stickster(a)gmail.com&gt;
> wrote:
> > > > Despite a posting to the lists earlier in the month[1] to check our
> > > > spin content, apparently no one caught the fact that there was no
> > > > Moblin spin among the release candidates.  At this point, we're going
> > > > to have to wait for mirrors to die down before we can consider
> > > > shipping a late copy.
> > >
> > > What's the latest update of this? If the mirrors are still too busy
> > > can we at least get the torrents up and propagating out with a link on
> > > the site to them?
> > >
> >
> > I'm working on creating the spin isos today.
>
>  Thanks, Jesse. 
WooHoo... nearly 2 days after the release and there's still no signs
of the Moblin spin. I'm not going to yell and scream like some would
about their spin but it would be nice to see some kind of update as to
why its still missing? I don't see that request as being unreasonable
after all I feel I put as much work into my spin and it can't help to
feel to me a little unfair and disappointed. If it was some other spin
all hell would have broken lose.

Peter

[Bruno Wolff III]

On Thu, May 27, 2010 at 07:37:48 +0100,
  Peter Robinson <pbrobinson(a)gmail.com&gt; wrote:
>  
>  WooHoo... nearly 2 days after the release and there's still no signs
>  of the Moblin spin. I'm not going to yell and scream like some would
>  about their spin but it would be nice to see some kind of update as to
>  why its still missing? I don't see that request as being unreasonable
>  after all I feel I put as much work into my spin and it can't help to
>  feel to me a little unfair and disappointed. If it was some other spin
>  all hell would have broken lose. 
I think that would depend on which spin. This isn't the first time something
bad has happened with a spin. I don't remember hell breaking loose in the
past.

Certainly it is unfortunate and disappointing that it happened.

[Rahul Sundaram]

On 05/27/2010 12:15 PM, Bruno Wolff III wrote:
>  On Thu, May 27, 2010 at 07:37:48 +0100,
>    Peter Robinson <pbrobinson(a)gmail.com&gt; wrote:
>    
> > WooHoo... nearly 2 days after the release and there's still no signs
> > of the Moblin spin. I'm not going to yell and scream like some would
> > about their spin but it would be nice to see some kind of update as to
> > why its still missing? I don't see that request as being unreasonable
> > after all I feel I put as much work into my spin and it can't help to
> > feel to me a little unfair and disappointed. If it was some other spin
> > all hell would have broken lose.
> >     
>  I think that would depend on which spin. This isn't the first time something
>  bad has happened with a spin. I don't remember hell breaking loose in the
>  past.
>
>  Certainly it is unfortunate and disappointing that it happened.
>     
I have long since advocated that spin owners be granted access in Fedora
infrastructure to compose their own spins and release engineering not
take this role.  The current method is simply not scaling well and we
have seen enough proof of that by now.
Rahul

[Kevin Fenzi]

ok. Let me start off by saying that I don't know if I have time/energy
to be the spins wrangler, but I will step up to do it if we can't find
any other interested parties. ;) 

(also note I am CC'ing John Poelstra on this email as he might have
input on the schedule I talk about below). 

So, I would propose we mail fedora-devel-announce with a Job posting
for 'spins wrangler' and see if we get an interested party stepping up
to do the job. ;) If we don't in a week, I will step up to do it. 

I would also like to propose some changes to the Spins process: 

Our current process deals with accepting spins and getting review and
approvals, etc. Once thats done it seems like we just say "ok, you're
done, thanks for the ks file" and expect everything to go great. I
don't think this is fair to our users or a good idea. I would like us
to add: 

- each spin for the cycle MUST have a contact email(s) for maintainers
  of that spin so we can contact them. (We could save this in a file in
  spins-kickstarts, etc). 

- I think we should gateway any git commits to the kickstart files to
  our mailing list for added visibility. 

- at some points in the development cycle we should REQUIRE a email or
  ACK from the maintainer of each spin. "My spin composes and has been
  tested by me or X others to work". This would include checking on
  sizes, etc. I think this points are up for discussion, but It might
  be nice to have them at: a week before Alpha, a week before Beta, at
  RC time before go/nogo, etc. 

- I agree we should have hard dates added of things that the spins
  wrangler and/or spins maintainers, and/or rel-eng need to do with
  spins. Here's my rough take: 

2010-06-01 - week after release, nightly composes start tracking
rawhide
2010-06-21 - All spins from previous cycle have updated wiki pages.
2010-07-06 - Deadline for submission of spins.
2010-07-27 - Feature Freeze/All spins known for release Wrangler
updates wiki/mails rel-eng/qa.
2010-07-27 - Spins wrangler requests bugzilla components for spins.
2010-08-03 - Alpha Change deadline - Go/No go from maintainers. 
2010-09-07 - Beta Change deadline - Go/No go from maintainers. 
2010-10-05 - Create spins-kickstarts package for release, branch git.
2010-10-05 - Spins ks Freeze. Only approved changes allowed. 
2010-10-14 - RC Composes - Go/No go from maintainers. 

This is missing "Spin maintainers submit text and screenshots for
spins.fedoraproject.org". When should that occur?

- I would like to have our next meeting (which would normally be
  tomorrow, but thats a holiday, so the following week) have a
  retrospective and see if there are other things we can improve. 

Thoughts?

kevin

[Bruno Wolff III]

On Sun, May 30, 2010 at 14:25:03 -0600,
  Kevin Fenzi <kevin(a)scrye.com&gt; wrote:
>  
>  - I would like to have our next meeting (which would normally be
>    tomorrow, but thats a holiday, so the following week) have a
>    retrospective and see if there are other things we can improve.  
Normally a Holiday would have been OK, but I have been sick the last three
days and am just starting to get some energy back.

I would like to see some more process documentation on the spin-kickstarts
stuff on when we want to branch, when repos should be changed and when new
rpms should be built and how to build the rpms from the git repo. (I am
already working on the latter, but didn't get much done this weekend.)

[John Poelstra]

Kevin Fenzi said the following on 05/30/2010 01:25 PM Pacific Time:
>  ok. Let me start off by saying that I don't know if I have
> time/energy
>  to be the spins wrangler, but I will step up to do it if we can't find
>  any other interested parties. ;)
>
>  (also note I am CC'ing John Poelstra on this email as he might have
>  input on the schedule I talk about below).
>
>  So, I would propose we mail fedora-devel-announce with a Job posting
>  for 'spins wrangler' and see if we get an interested party stepping up
>  to do the job. ;) If we don't in a week, I will step up to do it.
>
>  I would also like to propose some changes to the Spins process:
>
>  Our current process deals with accepting spins and getting review and
>  approvals, etc. Once thats done it seems like we just say "ok, you're
>  done, thanks for the ks file" and expect everything to go great. I
>  don't think this is fair to our users or a good idea. I would like us
>  to add:
>
>  - each spin for the cycle MUST have a contact email(s) for maintainers
>     of that spin so we can contact them. (We could save this in a file in
>     spins-kickstarts, etc).
>
>  - I think we should gateway any git commits to the kickstart files to
>     our mailing list for added visibility.
>
>  - at some points in the development cycle we should REQUIRE a email or
>     ACK from the maintainer of each spin. "My spin composes and has been
>     tested by me or X others to work". This would include checking on
>     sizes, etc. I think this points are up for discussion, but It might
>     be nice to have them at: a week before Alpha, a week before Beta, at
>     RC time before go/nogo, etc.
>
>  - I agree we should have hard dates added of things that the spins
>     wrangler and/or spins maintainers, and/or rel-eng need to do with
>     spins. Here's my rough take:
>
>  2010-06-01 - week after release, nightly composes start tracking
>  rawhide
>  2010-06-21 - All spins from previous cycle have updated wiki pages.
>  2010-07-06 - Deadline for submission of spins.
>  2010-07-27 - Feature Freeze/All spins known for release Wrangler
>  updates wiki/mails rel-eng/qa.
>  2010-07-27 - Spins wrangler requests bugzilla components for spins.
>  2010-08-03 - Alpha Change deadline - Go/No go from maintainers.
>  2010-09-07 - Beta Change deadline - Go/No go from maintainers.
>  2010-10-05 - Create spins-kickstarts package for release, branch git.
>  2010-10-05 - Spins ks Freeze. Only approved changes allowed.
>  2010-10-14 - RC Composes - Go/No go from maintainers.
>
>  This is missing "Spin maintainers submit text and screenshots for
>  spins.fedoraproject.org". When should that occur?
>
>  - I would like to have our next meeting (which would normally be
>     tomorrow, but thats a holiday, so the following week) have a
>     retrospective and see if there are other things we can improve.
>
>  Thoughts?
>
>  kevin 
I've created a formal version of this schedule here:
http://poelstra.fedorapeople.org/schedules/f-14/f-14-spins-tasks.html

I added additional milestones for more context.  Please let me know what 
you would like to change, add, or remove.

-------------------

I would highly recommend considering and adding the following to enable 
a smoother running Fedora 14 spins cycle compared to Fedora 13:

1) When testing will occur and for how long (through out)?

2) When you need Release Engineering to create the final spins?

3) When you plan to start test the final spins and for how long?

4) When Spins bittorrent files need to be created?

5) When final verification for final release day will be performed:
   --all the expected spins have been created
   --all the expected spins have been tested and work
   --bittorrent links work

6) Everything else that needs to be tracked and completed for a 
successful Fedora 14.
   --File design/artwork requests?

[Bruno Wolff III]

On Wed, Jun 09, 2010 at 09:46:26 -0700,
  John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  
>  I would highly recommend considering and adding the following to enable 
>  a smoother running Fedora 14 spins cycle compared to Fedora 13:
>  
>  2) When you need Release Engineering to create the final spins? 
And if we are going to do official spins for alpha and beta. We didn't for
F13 and I think that having people use the nightlies was better, but I
think that was a last minute decision and we really didn't consider it.

[Kevin Fenzi]

On Wed, 09 Jun 2010 09:46:26 -0700
John Poelstra <poelstra(a)redhat.com&gt; wrote:

>  I would highly recommend considering and adding the following to
>  enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
>  
>  1) When testing will occur and for how long (through out)? 
Yes. Thoughout. 

>  2) When you need Release Engineering to create the final spins?
Good question. 

I would say this would be the latest RC that gets accepted no?
So, 

2010-10-14 - RC composes start
The final spins would be created from whatever RC is accepted as final. 

>  3) When you plan to start test the final spins and for how long?
You mean the ones from the final freeze? Or the ones that rel-eng
creates?

>  4) When Spins bittorrent files need to be created? 
Those have typically been done when rel-eng is ready. Typically the
weekend before release, but it was monday (the day before release) this
last time. 

>  5) When final verification for final release day will be performed:
>     --all the expected spins have been created
>     --all the expected spins have been tested and work
>     --bittorrent links work 
Again, we should get rel-eng input. I would say the friday before
release? When RC is done? When bit links are made?

>  6) Everything else that needs to be tracked and completed for a 
>  successful Fedora 14.
>     --File design/artwork requests? 
Good questions. Artwork and spins.fedoraproject.org dates would be
great. I guess we need to ask websites and art about what dates should
be there?

kevin

[Kevin Fenzi]

On Wed, 9 Jun 2010 12:43:35 -0500
Bruno Wolff III <bruno(a)wolff.to&gt; wrote:

>  On Wed, Jun 09, 2010 at 09:46:26 -0700,
>    John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  > 
>  > I would highly recommend considering and adding the following to
>  > enable a smoother running Fedora 14 spins cycle compared to Fedora
>  > 13:
>  > 
>  > 2) When you need Release Engineering to create the final spins?
>  
>  And if we are going to do official spins for alpha and beta. We
>  didn't for F13 and I think that having people use the nightlies was
>  better, but I think that was a last minute decision and we really
>  didn't consider it.  
I'm fine with using the nightlies. 

If anyone would prefer more official images, speak up now? 
:) 

kevin

[John Poelstra]

Kevin Fenzi said the following on 06/10/2010 05:33 PM Pacific Time:
>  On Wed, 09 Jun 2010 09:46:26 -0700
>  John Poelstra<poelstra(a)redhat.com>  wrote:
>
> > I would highly recommend considering and adding the following to
> > enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
> >
> > 1) When testing will occur and for how long (through out)?
>
>  Yes. Thoughout. 
We might get better coverage if we explicitly specify when it it needs 
to happen vs. "all the time" :)

> > 2) When you need Release Engineering to create the final spins?
>
>  Good question.
>
>  I would say this would be the latest RC that gets accepted no?
>  So,
>
>  2010-10-14 - RC composes start
>  The final spins would be created from whatever RC is accepted as final. 
That is followed by the hand-off to QA on 2010-10-16 and then they have 
a week to qualify it.  This raises a couple of questions:

1) How long does it take to create the final spins? (releng)

2) How much time do you need to test them and when should testing start 
and end? (spins)

3) How much lead time does releng need to post the spins for GA, get the 
torrent files ready, etc?  IOW, when does the testing of the spins need 
to be completed by and spins that do not work dropped for Fedora 14?

> > 3) When you plan to start test the final spins and for how long?
>
>  You mean the ones from the final freeze? Or the ones that rel-eng
>  creates?
>  
Asked a different way above.

> > 4) When Spins bittorrent files need to be created?
>
>  Those have typically been done when rel-eng is ready. Typically the
>  weekend before release, but it was monday (the day before release) this
>  last time. 
I will put on the schedule for the Friday before GA.

> > 5) When final verification for final release day will be performed:
> >     --all the expected spins have been created
> >     --all the expected spins have been tested and work
> >     --bittorrent links work
>
>  Again, we should get rel-eng input. I would say the friday before
>  release? When RC is done? When bit links are made?
>
> > 6) Everything else that needs to be tracked and completed for a
> > successful Fedora 14.
> >     --File design/artwork requests?
>
>  Good questions. Artwork and spins.fedoraproject.org dates would be
>  great. I guess we need to ask websites and art about what dates should
>  be there?
>  
Yes, working with them to find out how much lead time they need is 
probably a great idea.  Then we can add some milestones to their 
schedule too so they can be thinking about it and preparing in advance.

John

[John Poelstra]

John Poelstra said the following on 06/16/2010 04:09 PM Pacific Time:
>  Kevin Fenzi said the following on 06/10/2010 05:33 PM Pacific Time:
> > On Wed, 09 Jun 2010 09:46:26 -0700
> > John Poelstra<poelstra(a)redhat.com>   wrote:
> >
> >> I would highly recommend considering and adding the following to
> >> enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
> >>
> >> 1) When testing will occur and for how long (through out)?
> >
> > Yes. Thoughout.
>
>  We might get better coverage if we explicitly specify when it it needs
>  to happen vs. "all the time" :)
>
> >
> >> 2) When you need Release Engineering to create the final spins?
> >
> > Good question.
> >
> > I would say this would be the latest RC that gets accepted no?
> > So,
> >
> > 2010-10-14 - RC composes start
> > The final spins would be created from whatever RC is accepted as final.
>
>  That is followed by the hand-off to QA on 2010-10-16 and then they have
>  a week to qualify it.  This raises a couple of questions:
>
>  1) How long does it take to create the final spins? (releng)
>
>  2) How much time do you need to test them and when should testing start
>  and end? (spins)
>
>  3) How much lead time does releng need to post the spins for GA, get the
>  torrent files ready, etc?  IOW, when does the testing of the spins need
>  to be completed by and spins that do not work dropped for Fedora 14?
>
> >
> >> 3) When you plan to start test the final spins and for how long?
> >
> > You mean the ones from the final freeze? Or the ones that rel-eng
> > creates?
> >
>
>  Asked a different way above.
>
> >> 4) When Spins bittorrent files need to be created?
> >
> > Those have typically been done when rel-eng is ready. Typically the
> > weekend before release, but it was monday (the day before release) this
> > last time.
>
>  I will put on the schedule for the Friday before GA.
>
> >
> >> 5) When final verification for final release day will be performed:
> >>      --all the expected spins have been created
> >>      --all the expected spins have been tested and work
> >>      --bittorrent links work
> >
> > Again, we should get rel-eng input. I would say the friday before
> > release? When RC is done? When bit links are made?
> >
> >> 6) Everything else that needs to be tracked and completed for a
> >> successful Fedora 14.
> >>      --File design/artwork requests?
> >
> > Good questions. Artwork and spins.fedoraproject.org dates would be
> > great. I guess we need to ask websites and art about what dates should
> > be there?
> >
>
>  Yes, working with them to find out how much lead time they need is
>  probably a great idea.  Then we can add some milestones to their
>  schedule too so they can be thinking about it and preparing in advance.
>
>  John 
This thread appears to have died.  Who is the leader of the spins sig 
these days and have regular meetings been re-established?

I'm concerned that if we do not set aside some time to reach a solid 
schedule in the next week we run the risk of repeating some of the same 
things that did not go well in Fedora 13.

John

[Bruno Wolff III]

On Wed, Jun 23, 2010 at 05:56:05 -0700,
  John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  
>  This thread appears to have died.  Who is the leader of the spins sig 
>  these days and have regular meetings been re-established? 
Kevin and I seem to be reluctantly sharing that. We are trying with the
meetings, but they aren't going well.

[Peter Robinson]

On Wed, Jun 23, 2010 at 2:18 PM, Bruno Wolff III <bruno(a)wolff.to&gt; wrote:
>  On Wed, Jun 23, 2010 at 05:56:05 -0700,
>   John Poelstra <poelstra(a)redhat.com&gt; wrote:
> >
> > This thread appears to have died.  Who is the leader of the spins sig
> > these days and have regular meetings been re-established?
>
>  Kevin and I seem to be reluctantly sharing that. We are trying with the
>  meetings, but they aren't going well. 
I apologise for my lack of response. I need to reply a number of the
spins email but the last 6 weeks have been crazy. I hope to be able to
more active involvement come July.

Peter

[Sebastian Dziallas]

On Wed, Jun 23, 2010 at 10:54 PM, Peter Robinson <pbrobinson(a)gmail.com&gt; wrote:
>  On Wed, Jun 23, 2010 at 2:18 PM, Bruno Wolff III
> <bruno(a)wolff.to&gt; wrote:
> > On Wed, Jun 23, 2010 at 05:56:05 -0700,
> >  John Poelstra <poelstra(a)redhat.com&gt; wrote:
> >>
> >> This thread appears to have died.  Who is the leader of the spins sig
> >> these days and have regular meetings been re-established?
> >
> > Kevin and I seem to be reluctantly sharing that. We are trying with the
> > meetings, but they aren't going well.
>
>  I apologise for my lack of response. I need to reply a number of the
>  spins email but the last 6 weeks have been crazy. I hope to be able to
>  more active involvement come July.
>
>  Peter 
Yeah, same here. I've been a slacker with regard to the meetings
lately. Thank you, Kevin and Bruno, for not letting this die. I'll try
to do a better job here, too.

--Sebastian

>  _______________________________________________
>  spins mailing list
>  spins(a)lists.fedoraproject.org
>  https://admin.fedoraproject.org/mailman/listinfo/spins

[Bruno Wolff III]

On Wed, Jun 16, 2010 at 16:09:00 -0700,
  John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  Kevin Fenzi said the following on 06/10/2010 05:33 PM Pacific Time:
>  > On Wed, 09 Jun 2010 09:46:26 -0700
>  > John Poelstra&lt;poelstra(a)redhat.com&gt;  wrote:
>  >
>  >> I would highly recommend considering and adding the following to
>  >> enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
>  >>
>  >> 1) When testing will occur and for how long (through out)?
>  >
>  > Yes. Thoughout.
>  
>  We might get better coverage if we explicitly specify when it it needs 
>  to happen vs. "all the time" :) 
I'd like to see extra covverage in the one week periods before the alpha,
beta and final development freezes. You kind of need to keep an eye on
things the whole time, as changes in other areas of Fedora can break your
spin in various ways. But the above times are good times to be doing a closer
look to see if problems have cropped up.

>  1) How long does it take to create the final spins? (releng) 
I don't know the answer to this one.

>  2) How much time do you need to test them and when should testing
> start 
>  and end? (spins) 
This is going to depend on the spin. I am not sure if all of the spins
owners have fast links. So it would be best to figure at least a day.
We haven't had mandatory confirmation required in the past. Is this something
you are thinking we should do for F14?

[John Poelstra]

Bruno Wolff III said the following on 06/30/2010 02:21 PM Pacific Time:
>  On Wed, Jun 16, 2010 at 16:09:00 -0700,
>     John Poelstra<poelstra(a)redhat.com>  wrote:
> > Kevin Fenzi said the following on 06/10/2010 05:33 PM Pacific Time:
> >> On Wed, 09 Jun 2010 09:46:26 -0700
> >> John Poelstra<poelstra(a)redhat.com>   wrote:
> >>
> >>> I would highly recommend considering and adding the following to
> >>> enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
> >>>
> >>> 1) When testing will occur and for how long (through out)?
> >>
> >> Yes. Thoughout.
> >
> > We might get better coverage if we explicitly specify when it it needs
> > to happen vs. "all the time" :)
>
>  I'd like to see extra covverage in the one week periods before the alpha,
>  beta and final development freezes. You kind of need to keep an eye on
>  things the whole time, as changes in other areas of Fedora can break your
>  spin in various ways. But the above times are good times to be doing a closer
>  look to see if problems have cropped up.
>
> > 1) How long does it take to create the final spins? (releng)
>
>  I don't know the answer to this one.
>  
Jesse--2 days to be safe?


> > 2) How much time do you need to test them and when should testing
> start
> > and end? (spins)
>
>  This is going to depend on the spin. I am not sure if all of the spins
>  owners have fast links. So it would be best to figure at least a day.
>  We haven't had mandatory confirmation required in the past. Is this something
>  you are thinking we should do for F14? 
I thought we should have had it for Fedora 12 :)

It has always bothered me we would ship something without clearly 
confirming and tracking whether the final product was explicitly tested.

In a perfect world I'd say a week to test.  It may not take that long to 
do, but some people might not get to it right away.

John

[Bruno Wolff III]

On Wed, Jun 30, 2010 at 14:41:14 -0700,
  John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  
>  It has always bothered me we would ship something without clearly
>  confirming and tracking whether the final product was explicitly
>  tested.
>  
>  In a perfect world I'd say a week to test.  It may not take that
>  long to do, but some people might not get to it right away. 
What is the fallback plan if an owner doesn't sign off on their spin?

Is someone else from spins supposed to do it, do we not publish a downloadable
iso for that release, delay publishing the iso, publish it but then chastise
the owner?

[John Poelstra]

Bruno Wolff III said the following on 06/30/2010 02:55 PM Pacific Time:
>  On Wed, Jun 30, 2010 at 14:41:14 -0700,
>     John Poelstra<poelstra(a)redhat.com>  wrote:
> >
> > It has always bothered me we would ship something without clearly
> > confirming and tracking whether the final product was explicitly
> > tested.
> >
> > In a perfect world I'd say a week to test.  It may not take that
> > long to do, but some people might not get to it right away.
>
>  What is the fallback plan if an owner doesn't sign off on their spin?
>
>  Is someone else from spins supposed to do it, do we not publish a downloadable
>  iso for that release, delay publishing the iso, publish it but then chastise
>  the owner? 
(removed extraneous CCs)

To me it is just like the feature process.  If the feature owner does 
not do their part (keep pages update, etc.) Fedora has no obligation to 
do theirs (advertise as part of the release, test, etc.).  So if the 
spin owner does not sign off that it has been tested, it doesn't get 
released for Fedora 14 and they can try again for Fedora 15.

We'll have regular schedule reminders, etc. so none of this will be a 
surprise.  I don't think the Spin SIG should have to cover for every 
spin owner that doesn't follow through.

I'm not looking to be overly harsh here.  There is definitely room to 
help each other out, give reminders, etc., but there comes a point where 
is someone says they are going to be the owner of something they have to 
do just that.  And the fact of the matter is you guys are already doing 
a lot already.

Thanks to you and Kevin for stepping up and helping to get things 
running smoothly here.

John

[Bruno Wolff III]

On Wed, Jun 30, 2010 at 15:08:43 -0700,
  John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  
>  To me it is just like the feature process.  If the feature owner does 
>  not do their part (keep pages update, etc.) Fedora has no obligation to 
>  do theirs (advertise as part of the release, test, etc.).  So if the 
>  spin owner does not sign off that it has been tested, it doesn't get 
>  released for Fedora 14 and they can try again for Fedora 15.
>  
>  We'll have regular schedule reminders, etc. so none of this will be a 
>  surprise.  I don't think the Spin SIG should have to cover for every 
>  spin owner that doesn't follow through. 
This will be a topic for the next meeting, but I expect we will recommend
requiring a confirmation that spin owners have tested the version of
GA version of their spin before GA day. If there are problems, they try
to work things out with releng to fix in time. (And maybe delay the release
of that particular spin.) If there is no confirmation, then the spin
is not released as an iso, though the ks file would still be available for
people that wanted to make their own.

We'll probably want an estimated start day and a deadline for the review.
At least one reminder close the the event, and another announcing the
change as well as fixing documentation.

[Bruno Wolff III]

I wanted to do a bit more with the spin-kickstarts package.
Currently the abridged schedule is:
 	Name  	Start  	End  	Duration
5. 	Alpha Change Deadline 	Tue 2010-08-03 	Tue 2010-08-03 	0
6. 	Alpha Public Availability 	Tue 2010-08-17 	Tue 2010-08-17 	0
7. 	Beta Change Deadline 	Tue 2010-09-07 	Tue 2010-09-07 	0
8. 	Beta Release Public Availability 	Tue 2010-09-21 	Tue 2010-09-21 	0
9. 	Spins kickstart package Freeze 	Tue 2010-10-05 	Tue 2010-10-05 	0
10. 	Create spins-kickstarts package for release and branch git 	Tue 2010-10-05 	Tue
2010-10-05 	0
11. 	Final Change Deadline 	Mon 2010-10-11 	Mon 2010-10-11 	0
12. 	Final (GA) Release 	Tue 2010-10-26 	Tue 2010-10-26 	0

I'd like to change 10 to:
10.    Branch spin-kickstarts and build package from new branch.

I'd like to add these steps:
5a.     Build spin-kickstarts package from master.
7a.     Build spin-kickstarts package from master.
12a.    If branched spin-kickstarts was changed build a new version for updates.

[John Poelstra]

Bruno Wolff III said the following on 06/30/2010 04:32 PM Pacific Time:
>  I wanted to do a bit more with the spin-kickstarts package.
>  Currently the abridged schedule is:
>    	Name  	Start  	End  	Duration
>  5. 	Alpha Change Deadline 	Tue 2010-08-03 	Tue 2010-08-03 	0
>  6. 	Alpha Public Availability 	Tue 2010-08-17 	Tue 2010-08-17 	0
>  7. 	Beta Change Deadline 	Tue 2010-09-07 	Tue 2010-09-07 	0
>  8. 	Beta Release Public Availability 	Tue 2010-09-21 	Tue 2010-09-21 	0
>  9. 	Spins kickstart package Freeze 	Tue 2010-10-05 	Tue 2010-10-05 	0
>  10. 	Create spins-kickstarts package for release and branch git 	Tue 2010-10-05 	Tue
> 2010-10-05 	0
>  11. 	Final Change Deadline 	Mon 2010-10-11 	Mon 2010-10-11 	0
>  12. 	Final (GA) Release 	Tue 2010-10-26 	Tue 2010-10-26 	0
>
>  I'd like to change 10 to:
>  10.    Branch spin-kickstarts and build package from new branch.
>
>  I'd like to add these steps:
>  5a.     Build spin-kickstarts package from master.
>  7a.     Build spin-kickstarts package from master.
>  12a.    If branched spin-kickstarts was changed build a new version for updates. 
Will do.  I'm doing a bunch of schedule updates targeted for Friday. 
I'll send a new link when they are updated.

John

[Paul W. Frields]

On Wed, Jun 30, 2010 at 06:24:34PM -0500, Bruno Wolff III wrote:
>  On Wed, Jun 30, 2010 at 15:08:43 -0700,
>    John Poelstra <poelstra(a)redhat.com&gt; wrote:
>  > 
>  > To me it is just like the feature process.  If the feature owner does 
>  > not do their part (keep pages update, etc.) Fedora has no obligation to 
>  > do theirs (advertise as part of the release, test, etc.).  So if the 
>  > spin owner does not sign off that it has been tested, it doesn't get 
>  > released for Fedora 14 and they can try again for Fedora 15.
>  > 
>  > We'll have regular schedule reminders, etc. so none of this will be a 
>  > surprise.  I don't think the Spin SIG should have to cover for every 
>  > spin owner that doesn't follow through.
>  
>  This will be a topic for the next meeting, but I expect we will recommend
>  requiring a confirmation that spin owners have tested the version of
>  GA version of their spin before GA day. If there are problems, they try
>  to work things out with releng to fix in time. (And maybe delay the release
>  of that particular spin.) If there is no confirmation, then the spin
>  is not released as an iso, though the ks file would still be available for
>  people that wanted to make their own.
>  
>  We'll probably want an estimated start day and a deadline for the review.
>  At least one reminder close the the event, and another announcing the
>  change as well as fixing documentation. 
IIRC Jesse Keating from the rel-eng team sent some reminders in the
F13 cycle about testing and inclusion.  It would make sense for
someone on the Spins SIG to consult with the rel-eng team to make sure
those start and deadline days jibe with the rest of the release
schedule.

Your description above that, regarding GA, sounds a lot like a
"go/no-go" for spins, which I think is a very good idea.  Whether or
not there is a chance to release later than that is a separate issue
and would also need to be discussed with the rel-eng team so that
everyone's expectations are clear before we get to that point.  I get
the feeling you're headed in that same direction, I just wanted to
call it out explicitly.

-- 
Paul W. Frields                                http://paul.frields.org/
  gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233  5906 ACDB C937 BD11 3717
  http://redhat.com/   -  -  -  -   http://pfrields.fedorapeople.org/
          Where open source multiplies: http://opensource.com

[John Poelstra]

Bruno Wolff III said the following on 06/30/2010 04:32 PM Pacific Time:
>  I wanted to do a bit more with the spin-kickstarts package.
>  Currently the abridged schedule is:
>    	Name  	Start  	End  	Duration
>  5. 	Alpha Change Deadline 	Tue 2010-08-03 	Tue 2010-08-03 	0
>  6. 	Alpha Public Availability 	Tue 2010-08-17 	Tue 2010-08-17 	0
>  7. 	Beta Change Deadline 	Tue 2010-09-07 	Tue 2010-09-07 	0
>  8. 	Beta Release Public Availability 	Tue 2010-09-21 	Tue 2010-09-21 	0
>  9. 	Spins kickstart package Freeze 	Tue 2010-10-05 	Tue 2010-10-05 	0
>  10. 	Create spins-kickstarts package for release and branch git 	Tue 2010-10-05 	Tue
> 2010-10-05 	0
>  11. 	Final Change Deadline 	Mon 2010-10-11 	Mon 2010-10-11 	0
>  12. 	Final (GA) Release 	Tue 2010-10-26 	Tue 2010-10-26 	0
>
>  I'd like to change 10 to:
>  10.    Branch spin-kickstarts and build package from new branch.
>
>  I'd like to add these steps:
>  5a.     Build spin-kickstarts package from master.
>  7a.     Build spin-kickstarts package from master.
>  12a.    If branched spin-kickstarts was changed build a new version for updates. 
http://poelstra.fedorapeople.org/schedules/f-14/f-14-spins-tasks.html

All added/changed. Let me know what else you need and how I can help.

Still to do:
1) Include final evaluation date
2) Include testing dates
3) Include official spins creation on the releng schedule once we hear 
back releng.

John

[Jesse Keating]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 6/30/10 2:41 PM, John Poelstra wrote:
>  Bruno Wolff III said the following on 06/30/2010 02:21 PM Pacific
> Time:
> > On Wed, Jun 16, 2010 at 16:09:00 -0700,
> >    John Poelstra<poelstra(a)redhat.com>  wrote:
> >> Kevin Fenzi said the following on 06/10/2010 05:33 PM Pacific Time:
> >>> On Wed, 09 Jun 2010 09:46:26 -0700
> >>> John Poelstra<poelstra(a)redhat.com>   wrote:
> >>>
> >>>> I would highly recommend considering and adding the following to
> >>>> enable a smoother running Fedora 14 spins cycle compared to Fedora 13:
> >>>>
> >>>> 1) When testing will occur and for how long (through out)?
> >>>
> >>> Yes. Thoughout.
> >>
> >> We might get better coverage if we explicitly specify when it it needs
> >> to happen vs. "all the time" :)
> >
> > I'd like to see extra covverage in the one week periods before the alpha,
> > beta and final development freezes. You kind of need to keep an eye on
> > things the whole time, as changes in other areas of Fedora can break your
> > spin in various ways. But the above times are good times to be doing a closer
> > look to see if problems have cropped up.
> >
> >> 1) How long does it take to create the final spins? (releng)
> >
> > I don't know the answer to this one.
> >
>  
>  Jesse--2 days to be safe? 
Yeah.  It largely depends on how many spins we do and the sizes of those
spins, but 2 days is a good window.

>  
>  
> >> 2) How much time do you need to test them and when should testing start
> >> and end? (spins)
> >
> > This is going to depend on the spin. I am not sure if all of the spins
> > owners have fast links. So it would be best to figure at least a day.
> > We haven't had mandatory confirmation required in the past. Is this something
> > you are thinking we should do for F14?
>  
>  I thought we should have had it for Fedora 12 :)
>  
>  It has always bothered me we would ship something without clearly 
>  confirming and tracking whether the final product was explicitly tested.
>  
>  In a perfect world I'd say a week to test.  It may not take that long to 
>  do, but some people might not get to it right away. 

- -- 
Jesse Keating
Fedora -- Freedom² is a feature!
identi.ca: http://identi.ca/jkeating
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (Darwin)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>
> iEYEARECAAYFAkwrwtIACgkQ4v2HLvE71NUQgQCfdizz0QELU82lOtKY51G9QbbQ
> 1sIAoLVDR++/3Ggo0C8GGd/NPRhuDGcO
> =+exL
> -----END PGP SIGNATURE-----