[Qemu-devel] Proposal for 2.9 release schedule

[Qemu-devel] Proposal for 2.9 release schedule

[Paolo Bonzini]

Considering that Easter is on April 16th, we'd probably want to have the
release before that date even in case of a slip.

On the other hand, the Christmas / New Year break here means that we'll
have to make the development time 1-2 week shorter in practice.

2016-02-21     2.9 soft freeze
2016-03-07     hard freeze / rc0
2016-03-28     rc3 (+3 weeks)
2016-04-04     rc4 or release
2016-04-11     release (if rc4)

One possibility is to make soft freeze happen a few days later.
Peter/Stefan, how did the experiment go with the new rules for soft
freeze?  Is it worth repeating it for 2.9 and would it make sense to
shorten soft freeze given the new rules?

Thanks,

Paolo

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 953 bytes --]

On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
> Considering that Easter is on April 16th, we'd probably want to have the
> release before that date even in case of a slip.
> 
> On the other hand, the Christmas / New Year break here means that we'll
> have to make the development time 1-2 week shorter in practice.
> 
> 2016-02-21     2.9 soft freeze
> 2016-03-07     hard freeze / rc0
> 2016-03-28     rc3 (+3 weeks)
> 2016-04-04     rc4 or release
> 2016-04-11     release (if rc4)
> 
> One possibility is to make soft freeze happen a few days later.
> Peter/Stefan, how did the experiment go with the new rules for soft
> freeze?  Is it worth repeating it for 2.9 and would it make sense to
> shorten soft freeze given the new rules?

I would shorten the soft freeze by 1 week.

Overall the 2.8 release went smoothly.  We got unlucky right at the end
with a release blocker but otherwise it was fine.

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Paolo Bonzini]

On 03/01/2017 16:53, Stefan Hajnoczi wrote:
> On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
>> Considering that Easter is on April 16th, we'd probably want to have the
>> release before that date even in case of a slip.
>>
>> On the other hand, the Christmas / New Year break here means that we'll
>> have to make the development time 1-2 week shorter in practice.
>>
>> 2016-02-21     2.9 soft freeze
>> 2016-03-07     hard freeze / rc0
>> 2016-03-28     rc3 (+3 weeks)
>> 2016-04-04     rc4 or release
>> 2016-04-11     release (if rc4)
>>
>> One possibility is to make soft freeze happen a few days later.
>> Peter/Stefan, how did the experiment go with the new rules for soft
>> freeze?  Is it worth repeating it for 2.9 and would it make sense to
>> shorten soft freeze given the new rules?
> 
> I would shorten the soft freeze by 1 week.
> 
> Overall the 2.8 release went smoothly.  We got unlucky right at the end
> with a release blocker but otherwise it was fine.

Then what about soft freeze on 2016-02-28?

Thanks,

Paolo

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 1213 bytes --]

On Tue, Jan 03, 2017 at 05:06:13PM +0100, Paolo Bonzini wrote:
> On 03/01/2017 16:53, Stefan Hajnoczi wrote:
> > On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
> >> Considering that Easter is on April 16th, we'd probably want to have the
> >> release before that date even in case of a slip.
> >>
> >> On the other hand, the Christmas / New Year break here means that we'll
> >> have to make the development time 1-2 week shorter in practice.
> >>
> >> 2016-02-21     2.9 soft freeze
> >> 2016-03-07     hard freeze / rc0
> >> 2016-03-28     rc3 (+3 weeks)
> >> 2016-04-04     rc4 or release
> >> 2016-04-11     release (if rc4)
> >>
> >> One possibility is to make soft freeze happen a few days later.
> >> Peter/Stefan, how did the experiment go with the new rules for soft
> >> freeze?  Is it worth repeating it for 2.9 and would it make sense to
> >> shorten soft freeze given the new rules?
> > 
> > I would shorten the soft freeze by 1 week.
> > 
> > Overall the 2.8 release went smoothly.  We got unlucky right at the end
> > with a release blocker but otherwise it was fine.
> 
> Then what about soft freeze on 2016-02-28?

Sounds good to me.  Peter?

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Peter Maydell]

On 4 January 2017 at 14:51, Stefan Hajnoczi <stefanha@redhat.com> wrote:
> On Tue, Jan 03, 2017 at 05:06:13PM +0100, Paolo Bonzini wrote:
>> On 03/01/2017 16:53, Stefan Hajnoczi wrote:
>> > On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
>> >> Considering that Easter is on April 16th, we'd probably want to have the
>> >> release before that date even in case of a slip.
>> >>
>> >> On the other hand, the Christmas / New Year break here means that we'll
>> >> have to make the development time 1-2 week shorter in practice.
>> >>
>> >> 2016-02-21     2.9 soft freeze
>> >> 2016-03-07     hard freeze / rc0
>> >> 2016-03-28     rc3 (+3 weeks)
>> >> 2016-04-04     rc4 or release
>> >> 2016-04-11     release (if rc4)
>> >>
>> >> One possibility is to make soft freeze happen a few days later.
>> >> Peter/Stefan, how did the experiment go with the new rules for soft
>> >> freeze?  Is it worth repeating it for 2.9 and would it make sense to
>> >> shorten soft freeze given the new rules?
>> >
>> > I would shorten the soft freeze by 1 week.
>> >
>> > Overall the 2.8 release went smoothly.  We got unlucky right at the end
>> > with a release blocker but otherwise it was fine.
>>
>> Then what about soft freeze on 2016-02-28?
>
> Sounds good to me.  Peter?

Are we retaining the "make sure you have your pull requests on the list
by the softfreeze date" rule this time around?

thanks
-- PMM

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 1619 bytes --]

On Fri, Jan 06, 2017 at 03:12:28PM +0000, Peter Maydell wrote:
> On 4 January 2017 at 14:51, Stefan Hajnoczi <stefanha@redhat.com> wrote:
> > On Tue, Jan 03, 2017 at 05:06:13PM +0100, Paolo Bonzini wrote:
> >> On 03/01/2017 16:53, Stefan Hajnoczi wrote:
> >> > On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
> >> >> Considering that Easter is on April 16th, we'd probably want to have the
> >> >> release before that date even in case of a slip.
> >> >>
> >> >> On the other hand, the Christmas / New Year break here means that we'll
> >> >> have to make the development time 1-2 week shorter in practice.
> >> >>
> >> >> 2016-02-21     2.9 soft freeze
> >> >> 2016-03-07     hard freeze / rc0
> >> >> 2016-03-28     rc3 (+3 weeks)
> >> >> 2016-04-04     rc4 or release
> >> >> 2016-04-11     release (if rc4)
> >> >>
> >> >> One possibility is to make soft freeze happen a few days later.
> >> >> Peter/Stefan, how did the experiment go with the new rules for soft
> >> >> freeze?  Is it worth repeating it for 2.9 and would it make sense to
> >> >> shorten soft freeze given the new rules?
> >> >
> >> > I would shorten the soft freeze by 1 week.
> >> >
> >> > Overall the 2.8 release went smoothly.  We got unlucky right at the end
> >> > with a release blocker but otherwise it was fine.
> >>
> >> Then what about soft freeze on 2016-02-28?
> >
> > Sounds good to me.  Peter?
> 
> Are we retaining the "make sure you have your pull requests on the list
> by the softfreeze date" rule this time around?

I hope so.  It helps keep the freeze time bounded.

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Peter Maydell]

On 9 January 2017 at 10:41, Stefan Hajnoczi <stefanha@redhat.com> wrote:
> On Fri, Jan 06, 2017 at 03:12:28PM +0000, Peter Maydell wrote:
>> On 4 January 2017 at 14:51, Stefan Hajnoczi <stefanha@redhat.com> wrote:
>> > On Tue, Jan 03, 2017 at 05:06:13PM +0100, Paolo Bonzini wrote:
>> >> On 03/01/2017 16:53, Stefan Hajnoczi wrote:
>> >> > On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
>> >> >> Considering that Easter is on April 16th, we'd probably want to have the
>> >> >> release before that date even in case of a slip.
>> >> >>
>> >> >> On the other hand, the Christmas / New Year break here means that we'll
>> >> >> have to make the development time 1-2 week shorter in practice.
>> >> >>
>> >> >> 2016-02-21     2.9 soft freeze
>> >> >> 2016-03-07     hard freeze / rc0
>> >> >> 2016-03-28     rc3 (+3 weeks)
>> >> >> 2016-04-04     rc4 or release
>> >> >> 2016-04-11     release (if rc4)
>> >> >>
>> >> >> One possibility is to make soft freeze happen a few days later.
>> >> >> Peter/Stefan, how did the experiment go with the new rules for soft
>> >> >> freeze?  Is it worth repeating it for 2.9 and would it make sense to
>> >> >> shorten soft freeze given the new rules?
>> >> >
>> >> > I would shorten the soft freeze by 1 week.
>> >> >
>> >> > Overall the 2.8 release went smoothly.  We got unlucky right at the end
>> >> > with a release blocker but otherwise it was fine.
>> >>
>> >> Then what about soft freeze on 2016-02-28?
>> >
>> > Sounds good to me.  Peter?
>>
>> Are we retaining the "make sure you have your pull requests on the list
>> by the softfreeze date" rule this time around?
>
> I hope so.  It helps keep the freeze time bounded.

OK. The dates above work ok for me, so I've updated the wiki:
http://wiki.qemu.org/Planning/2.9

If we're going to standardize on the new softfreeze definition we should
update http://wiki.qemu.org/Planning/SoftFeatureFreeze I guess.

thanks
-- PMM

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Paolo Bonzini]

On 09/01/2017 12:11, Peter Maydell wrote:
> On 9 January 2017 at 10:41, Stefan Hajnoczi <stefanha@redhat.com> wrote:
>> On Fri, Jan 06, 2017 at 03:12:28PM +0000, Peter Maydell wrote:
>>> On 4 January 2017 at 14:51, Stefan Hajnoczi <stefanha@redhat.com> wrote:
>>>> On Tue, Jan 03, 2017 at 05:06:13PM +0100, Paolo Bonzini wrote:
>>>>> On 03/01/2017 16:53, Stefan Hajnoczi wrote:
>>>>>> On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
>>>>>>> Considering that Easter is on April 16th, we'd probably want to have the
>>>>>>> release before that date even in case of a slip.
>>>>>>>
>>>>>>> On the other hand, the Christmas / New Year break here means that we'll
>>>>>>> have to make the development time 1-2 week shorter in practice.
>>>>>>>
>>>>>>> 2016-02-21     2.9 soft freeze
>>>>>>> 2016-03-07     hard freeze / rc0
>>>>>>> 2016-03-28     rc3 (+3 weeks)
>>>>>>> 2016-04-04     rc4 or release
>>>>>>> 2016-04-11     release (if rc4)
>>>>>>>
>>>>>>> One possibility is to make soft freeze happen a few days later.
>>>>>>> Peter/Stefan, how did the experiment go with the new rules for soft
>>>>>>> freeze?  Is it worth repeating it for 2.9 and would it make sense to
>>>>>>> shorten soft freeze given the new rules?
>>>>>>
>>>>>> I would shorten the soft freeze by 1 week.
>>>>>>
>>>>>> Overall the 2.8 release went smoothly.  We got unlucky right at the end
>>>>>> with a release blocker but otherwise it was fine.
>>>>>
>>>>> Then what about soft freeze on 2016-02-28?
>>>>
>>>> Sounds good to me.  Peter?
>>>
>>> Are we retaining the "make sure you have your pull requests on the list
>>> by the softfreeze date" rule this time around?
>>
>> I hope so.  It helps keep the freeze time bounded.
> 
> OK. The dates above work ok for me, so I've updated the wiki:
> http://wiki.qemu.org/Planning/2.9
> 
> If we're going to standardize on the new softfreeze definition we should
> update http://wiki.qemu.org/Planning/SoftFeatureFreeze I guess.

Done, any help with the wording is welcome of course.

Paolo

Re: [Qemu-devel] Proposal for 2.9 release schedule

[Peter Maydell]

On 3 January 2017 at 16:53, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> On Fri, Dec 23, 2016 at 03:15:58PM +0100, Paolo Bonzini wrote:
>> Considering that Easter is on April 16th, we'd probably want to have the
>> release before that date even in case of a slip.
>>
>> On the other hand, the Christmas / New Year break here means that we'll
>> have to make the development time 1-2 week shorter in practice.
>>
>> 2016-02-21     2.9 soft freeze
>> 2016-03-07     hard freeze / rc0
>> 2016-03-28     rc3 (+3 weeks)
>> 2016-04-04     rc4 or release
>> 2016-04-11     release (if rc4)

[these should all have been 2017, heh]

>> One possibility is to make soft freeze happen a few days later.
>> Peter/Stefan, how did the experiment go with the new rules for soft
>> freeze?  Is it worth repeating it for 2.9 and would it make sense to
>> shorten soft freeze given the new rules?
>
> I would shorten the soft freeze by 1 week.

I hadn't noticed that this meant that we had a 1 week softfreeze
period. This completely failed, partly because I didn't think
we'd gone down to just 1 week, partly because I was away at a
conference last week when rc0 was theoretically due, and
partly because of the enormous pile of merges that needed
to be done -- I only finished the merge work on the Saturday.

I plan to tag rc0 tomorrow (the 14th) and push all the other
dates out by a week accordingly.

thanks
-- PMM