Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

First pass at 1.8 schedule #305

Merged
merged 1 commit into from
Jul 21, 2017
Merged

Conversation

grodrigues3
Copy link
Contributor

Approximate timeline and schedule for the Kubernetes 1.8 release

@kubernetes/sig-release-misc

@k8s-ci-robot k8s-ci-robot added sig/release Categorizes an issue or PR as relevant to SIG Release. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. labels Jun 22, 2017
@@ -0,0 +1,13 @@
### Release Team absence tracker
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we really need this file?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Probably not, I can delete

Copy link
Member

@luxas luxas left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A couple of comments...
Otherwise LGTM

* Ensure docs and release notes are written
* Identify all features going into the release, and make sure alpha, beta, ga is marked in features repo
- Sept 6 (Weds): v1.8.0-beta.0
- Jun 7 (Weds): v1.8.0-beta.1 cut
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should be Sept 13?

* Focus on bugfix, test flakes and stabilization
* Ensure docs and release notes are written
* Identify all features going into the release, and make sure alpha, beta, ga is marked in features repo
- Sept 6 (Weds): v1.8.0-beta.0
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Clarify that the release-1.8 branch is created at this point and v1.9.0-alpha.0 is created

* Begin regular fast-forwards (at least daily to keep CI up-to-date).

### Sept 1 (Fri) - Sept 14 (Thurs)
- Sept 1 Begin Feature Freeze
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Be consistent and say Code Freeze to be clear that this isn't the Features Repo Freeze

have an associated issue in the features repo
- Aug 9 (Weds): v1.8.0-alpha.3
- Aug 23 (Weds): v1.8.0-alpha.4
- Sept 6 (Weds): v1.8.0-beta.0
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This seems misplaced

- Release 1.8 alphas every 2 weeks
- July 12th (Weds): v1.8.0-alpha.1
- July 26th (Weds): v1.8.0-alpha.2
- Aug 1 (Tues): features repo freeze: all features going into release should
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sometimes 12th, 26th, etc. sometimes 1, 9, 23.
Please be consistent

- July 5 (Weds) coding start (7w)
- July 12th (Weds): v1.8.0-alpha.1
- July 26th (Weds): v1.8.0-alpha.2
- Aug 1 (Tues): features repo freeze
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

suggest, just say "Feature Freeze" here and define later down in the doc.
The term has been used for the past 5-ish releases so it's not new anyway

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1



# Key features
[Feature tracking spreadsheet
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Spreadsheet link to be provided by @idvoretskyi

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Resolved.

@pwittrock
Copy link
Member

/assign pwittrock


### July 5 - Sept 1
- 8 week coding period
- Release 1.8 alphas every 2 weeks
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i suggest we require all release tests to be passing to cut an alpha

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agreed. The more the alpha releases are like treated like real releases, the better.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To clarify:

  • By "all release tests", do you mean all the jobs displayed on the release-master-blocking testgrid dashboard? AFAIK this what kubernetes/release uses
  • By "be passing" do you mean their most-recent "Overall" results are green? Some noise I've heard in the past: human explanations for why we could ignore red results in some mostly green suites; a practice of waiting for "3 in a row" prior to calling a final release good

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

By "all release tests", do you mean all the jobs displayed on the release-master-blocking testgrid dashboard? AFAIK this what kubernetes/release uses

I think this makes sense at least until we can get better ownership of non-blocking suites. If we do include others at this point we should explicitly list them since the status quo is currently having some bad jobs. This would make sense for important suites that are likely non-blocking for test duration reasons like serial and soak.

By "be passing" do you mean their most-recent "Overall" results are green? Some noise I've heard in the past: human explanations for why we could ignore red results in some mostly green suites; a practice of waiting for "3 in a row" prior to calling a final release good

The "3 in a row" is the standard I've seen for the past few releases but it's generally been up to human discretion. We should decide on an exact criteria and ideally have it machine enforced.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe we need to look at overall test stability when it comes to making the call. Philosophically speaking, a disruptive release is far worse than a disruptive wait.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: don't cut alpha until master-blocking-tests are passing
Outstanding Question: GKE tests should be release blocking? Failing 25+% clouds the release signal.

- Aug 23rd (Weds): v1.8.0-alpha.4

### Sept 1 (Fri) - Sept 14 (Thurs)
- Sept 1st Begin Code Freeze
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should this be called 'Code Slush' or 'Code Freeze'?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Both terms have been used in the past...

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's pick one and be consistent. I have no strong opinions, so I consulted a search of kubernetes-dev. "Code freeze" returned ~58 results, "code slush" returned ~9, so maybe we should just stick with "code freeze" even if it doesn't strictly represent what happens during this period.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 for freeze

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am +1 for freeze, but wanted to point out that "code slush" usually implies some changes can/will continue to be made at the discretion of the release team. "Code freeze" from an auditing perspective means that zero code changes above those required for the release mechanics will occur. If we look at code freezes from the past, it might look more slushy than hard frozen. If we want to introduce a period of time were select changes were possible, we could use slush for that to differentiate.

* v1.9.0-alpha.0 is cut
* Test-infra lead to set up CI for branch.
* Begin regular fast-forwards (at least daily to keep CI up-to-date).
- Sept 13th (Weds): End of Code Slush
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I was thinking we should shoot for making code slush just 1 week. If the tests are all passing, and the issues are auto-managed, this should be possible.

Copy link
Member

@pwittrock pwittrock Jun 30, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should only accept cherry picks for release blockers starting 1 week after code slush starts.


### Sept 15 - Sept 27
- Sept 18th (Mon): v1.9.0-alpha.1
- Sept 20th (Weds): v1.8.release-candidate
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is effectively saying, have all release blocking issues resolved 1 week before we release. Historically, we have never come close to this, or even been able to resolve release blocking issues by the actual release date. I suggest either scrap the release candidate until we are able to consistently meet our release timeline.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the issue is this hasn't been treated as the actual release date in the past so we always blow right through it. I do think there's value in trying to target this because we still haven't released on time and this gives a buffer and the way we backload code changes means hitting any specific date is unlikely.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Isn't the release candidate supposed to provide a preview for external consumer validation, and other out-of-band actions? If it's not as close as possible to the actual release, I am not sure what value it has to the community. If it's a dry run for the actual release, I am all for it. If it's just "one more thing" to worry about at release crunch time with no demonstrable value, then it should be obsoleted.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We need to to treat RC date as the release date if we want the RC to be meaningful, which is something we haven't done well yet.

I'm in favor of keeping the RC but only if we commit to a mindset that Sept. 20 is when we're releasing.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: keep it but drive harder to

  • have all tests passing (blocking, upgrades, skew)

* Only accept cherrypicks for release-blockers.
* Announce on mailing lists, Twitter, etc. to ask people to build and test the RC and submit feedback
* Perhaps managed k8s providers make rc.1 available on early access channels.
* More RCs as needed to verify release-blocking fixes.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We need to have something about release notes and docs on here. When they are started, and what the dates for milestones are. I suggest release notes for completed issues must be done by SIGs 1 week before the release date.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

By now, we're 1 week out from release. From a release notes perspective, I'm in full agreement we should basically have everything documented, all we should be changing by now are structural edits or last minute Known Issues. We should have enough in place that the release branch manager could cut a copy of CHANGELOG.md with all the release notes incorporated (assuming that's where they still land)

IMO what should not be happening by now is the addition of features or large structural changes in the release notes. I feel like I saw both happen the final week of the previous release.

From a docs perspective, I would hope we're basically at a live copy of all of the docs, with grammar edits and spell checking at this point. Historically that has not been the case though. What is reasonable to expect by now? @kubernetes/sig-docs-maintainers

Copy link
Contributor

@chenopis chenopis Jul 6, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Honestly, at 1 week out for 1.7, half the 22 docs were merged and good to go. A couple days later (that Fri), we were at 80%. Even the day of the original release date, we were scrambling to get a couple docs reviewed and merged because one snuck in last minute and another one hadn't been completed and ready for review until that day. It follows sort of an exponential decay pattern. We can set the deadlines earlier, but there will still be a few stragglers at the end.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Resolved.

## Operational Changes
1. For the 1.8 release, the [release team](https://github.com/kubernetes/features/blob/master/release-1.8/release_team.md)
will use the following procedure to identify release blocking issues
1. Any issues listed in the [v1.8 milestone](https://github.com/kubernetes/kubernetes/issues?utf8=%E2%9C%93&q=is%3Aissue%20is%3Aopen%20milestone%3Av1.8)
Copy link
Member

@pwittrock pwittrock Jun 30, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FYI, my proposal is that issues are automatically moved out of the milestone unless they are labeled critical-urgent and have a SIG owner.

See link here
kubernetes/community#752

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let the 1.8 release team decide on this

milestones **or the release will not ship**

# Contact us
- [via slack](https://kubernetes.slack.com/messages/k8s-release/)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sig-release?

| Test Infra | | | |
| Docs || |
| Bugs | | | |
| CI Signal | | | |
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For 1.8 I think we should merge CI Signal and Upgrade Testing, but the role should start immediately and require keeping the tests passing throughout the release.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 for having roles start immediately

My only concern with merging would be spreading leads thin.

- Sept 6th (Weds): v1.8.0-beta.0
* Branch Manager to create v1.8 release branch.
* v1.9.0-alpha.0 is cut
* Test-infra lead to set up CI for branch.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Note that we will still use the master tests as the source of truth until the end of code slush, since the master tests are more up-to-date

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there a case where this practice could lead to a false green or red? Just wondering if there are any risks in this.

@@ -0,0 +1,12 @@
| **Role** | **Name** (**GitHub/Slack ID**) | **Shadow Name(s) (GitHub/Slack ID), ...**
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is the "Shadow role" still actual? For 1.7 we had only Lead release manager shadowed.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Personally I would like to see it remain present. I can shadow someone this release.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 I like how it provides a lower commitment avenue to be involved in a release

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be moot if we don't find someone to fill it.

@jdumars
Copy link
Member

jdumars commented Jul 5, 2017

With the requested changes, LGTM.

Copy link
Member

@spiffxp spiffxp left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd like to see requirements for docs spelled out a bit more, and some explicit definition of passing tests. Couple other tiny nits.

* Ensure docs and release notes are written
* Identify all features going into the release, and make sure alpha, beta, ga is marked in features repo
- Sept 6th (Weds): v1.8.0-beta.0
* Branch Manager to create v1.8 release branch.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

release-1.8 branch

* Identify all features going into the release, and make sure alpha, beta, ga is marked in features repo
- Sept 6th (Weds): v1.8.0-beta.0
* Branch Manager to create v1.8 release branch.
* v1.9.0-alpha.0 is cut
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

kubernetes/release#342 @timothysc does this schedule adequately address the concern you had here?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@spiffxp Yes, this is much better than v1.7, where we cut the branch basically the same day as the code freeze.
As we know there's a lot of movement when everyone is trying to get their thing into the next release, the SQ was 50-ish deep and things flaked violently.

Cutting the branch that day or the next wasn't the best thing to do, as folks had to fast-forward it very frequently in order to keep things somewhat green. This combined with all the other churn wasn't ideal. On top it this, it also made things harder for consumers of release artifacts like kubeadm.

Waiting a couple of days before cutting the branch/beta.0/alpha.0 makes a lot of sense, thanks @grodrigues3


### July 5 - Sept 1
- 8 week coding period
- Release 1.8 alphas every 2 weeks
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To clarify:

  • By "all release tests", do you mean all the jobs displayed on the release-master-blocking testgrid dashboard? AFAIK this what kubernetes/release uses
  • By "be passing" do you mean their most-recent "Overall" results are green? Some noise I've heard in the past: human explanations for why we could ignore red results in some mostly green suites; a practice of waiting for "3 in a row" prior to calling a final release good

- Aug 23rd (Weds): v1.8.0-alpha.4

### Sept 1 (Fri) - Sept 14 (Thurs)
- Sept 1st Begin Code Freeze
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's pick one and be consistent. I have no strong opinions, so I consulted a search of kubernetes-dev. "Code freeze" returned ~58 results, "code slush" returned ~9, so maybe we should just stick with "code freeze" even if it doesn't strictly represent what happens during this period.

- Sept 13th (Weds): End of Code Slush
* v1.8.0-beta.1 cut
* Final fast-forward of release branch.
* After this, all changes for the release must now be cherrypicked in batches by branch
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@pwittrock I believe at this point, we no longer look at master tests, but at release-branch tests for our CI signal, right?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@spiffxp yup, after we start cherrypicking master no longer duplicates CI

* Add milestone restriction on submit queue.
* Community Feature Burndown Meetings held two or three times until release week (then every day). For those interested in joining please join [Kubernetes Milestone Burndown Group](https://groups.google.com/forum/#!forum/kubernetes-milestone-burndown)
* Focus on bugfix, test flakes and stabilization
* Ensure docs and release notes are written
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@pwittrock so we mention release notes and docs up here at the start of code freeze, but agree we're not spelling out enough

In terms of release notes, we should now have enough to collate one-line descriptions of high-level features that are being added, deprecations or removals that are happening, and improvements that are being made (eg: scrape the one-line release description from features repo issues, or copy-pasta from a manually managed spreadsheet)

This should be enough to answer "but why should I care about the beta?" with a bite-sized blog or e-mail of "you can try out X, we've made Y way more usable, and you can can get a jump on the deprecation of Z"

In terms of docs, I'm guessing by now there should be enough content for a tech review? @kubernetes/sig-docs-maintainers

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@spiffxp this all seems spot on. Those three criteria must be satisfied. We should have a document go/no-go date set as well. Documentation (the one-liner) should be a blocker for matriculation IMO since the burden is low and the risk of end-user "feature discovery" is very high. Without this documentation, we lose our most valuable asset: transparency.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

how would we feel about making "Release note deadline" two weeks after feature freeze?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 @calebamiles. But I'd propose to set this deadline for docs as well, and at the same time, when the code freeze should happen.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it makes sense collect the full release notes at the time of code freeze.

Collecting and announcing what features are gonna be in the release on a high level should be done as soon as possible after the feature freeze. It will serve as the base for the release notes, which will be enhanced at the time of the code freeze with important PRs, action required/deprecation/etc items etc.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@calebamiles I am for that being when they are finalized, but not necessarily due. I acknowledge that the burden of getting the feature AND docs over the line at the same time can be difficult. But, SIGs should have that as part of the planned workstream as much as the feature itself. Would a one-week grace period serve the purpose? After that, it would be time to have discussions with SIGs and individuals to ensure it gets done.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: Relnotes for feature due after feature freeze (see dates above)
Decision: Docs out (must be a PR) for tech review by one week after code freeze. A week after that should be lgtm'd and ready for merge.

* Test-infra lead to set up CI for branch.
* Begin regular fast-forwards (at least daily to keep CI up-to-date).
- Sept 13th (Weds): End of Code Slush
* v1.8.0-beta.1 cut
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We've just lifted code freeze, cut our second beta, and are one full business week and change out from our first pass at docs and release notes. What is reasonable progress to expect by now? @kubernetes/sig-docs-maintainers

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm curious about this as well.

* Only accept cherrypicks for release-blockers.
* Announce on mailing lists, Twitter, etc. to ask people to build and test the RC and submit feedback
* Perhaps managed k8s providers make rc.1 available on early access channels.
* More RCs as needed to verify release-blocking fixes.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

By now, we're 1 week out from release. From a release notes perspective, I'm in full agreement we should basically have everything documented, all we should be changing by now are structural edits or last minute Known Issues. We should have enough in place that the release branch manager could cut a copy of CHANGELOG.md with all the release notes incorporated (assuming that's where they still land)

IMO what should not be happening by now is the addition of features or large structural changes in the release notes. I feel like I saw both happen the final week of the previous release.

From a docs perspective, I would hope we're basically at a live copy of all of the docs, with grammar edits and spell checking at this point. Historically that has not been the case though. What is reasonable to expect by now? @kubernetes/sig-docs-maintainers

@@ -0,0 +1,12 @@
| **Role** | **Name** (**GitHub/Slack ID**) | **Shadow Name(s) (GitHub/Slack ID), ...**
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Personally I would like to see it remain present. I can shadow someone this release.


# Contact us
- [via slack](https://kubernetes.slack.com/messages/k8s-release/)
- [via email](mailto:kubernetes-release@googlegroups.com)
Copy link
Member

@jdumars jdumars Jul 20, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Contributor

@calebamiles calebamiles left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We'll need to add in the dates for the release notes and docs but it all looks reasonable to me


### July 5 - Sept 1
- 8 week coding period
- Release 1.8 alphas every 2 weeks
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: don't cut alpha until master-blocking-tests are passing
Outstanding Question: GKE tests should be release blocking? Failing 25+% clouds the release signal.

- July 26th (Weds): v1.8.0-alpha.2
- Aug 1 (Tues): Feature Freeze
- Aug 9 (Weds): v1.8.0-alpha.3
- Aug 23 (Weds): v1.8.0-alpha.4
Copy link
Contributor Author

@grodrigues3 grodrigues3 Jul 21, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Aug 8: release notes draft for features due
Aug 15: hard deadline for feature related relnotes

  • Consequence: your feature not included in release notes

* Add milestone restriction on submit queue.
* Community Feature Burndown Meetings held two or three times until release week (then every day). For those interested in joining please join [Kubernetes Milestone Burndown Group](https://groups.google.com/forum/#!forum/kubernetes-milestone-burndown)
* Focus on bugfix, test flakes and stabilization
* Ensure docs and release notes are written
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: Relnotes for feature due after feature freeze (see dates above)
Decision: Docs out (must be a PR) for tech review by one week after code freeze. A week after that should be lgtm'd and ready for merge.


### Sept 15 - Sept 27
- Sept 18th (Mon): v1.9.0-alpha.1
- Sept 20th (Weds): v1.8.release-candidate
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Decision: keep it but drive harder to

  • have all tests passing (blocking, upgrades, skew)

* Only accept cherrypicks for release-blockers.
* Announce on mailing lists, Twitter, etc. to ask people to build and test the RC and submit feedback
* Perhaps managed k8s providers make rc.1 available on early access channels.
* More RCs as needed to verify release-blocking fixes.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Resolved.



# Key features
[Feature tracking spreadsheet
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Resolved.

## Operational Changes
1. For the 1.8 release, the [release team](https://github.com/kubernetes/features/blob/master/release-1.8/release_team.md)
will use the following procedure to identify release blocking issues
1. Any issues listed in the [v1.8 milestone](https://github.com/kubernetes/kubernetes/issues?utf8=%E2%9C%93&q=is%3Aissue%20is%3Aopen%20milestone%3Av1.8)
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let the 1.8 release team decide on this

@grodrigues3
Copy link
Contributor Author

Merging after the 07/21 scheduling finalization meeting

@grodrigues3 grodrigues3 merged commit 27fe14a into kubernetes:master Jul 21, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. sig/release Categorizes an issue or PR as relevant to SIG Release.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet