testgrid should show the new build SHA in upgrade test

[caesarxuchao]

e.g., https://k8s-testgrid.appspot.com/1.6-1.7-upgrade

The UI shows the SHA of the old image. It's more useful if the UI instead shows the SHA of the image the cluster is upgraded to, because that's what the release team cares about.

cc @krzyzacy @ixdy

[ixdy]

FWIW Gubernator used to show both versions in the metadata at the top, but in some of the job config and test-infra refactoring, it seems like this functionality was lost.

[fejta]

Update this method to include both versions in the metadata.json, which will result in it getting exposed to both gubernator and testgrid
https://github.com/kubernetes/test-infra/blob/master/kubetest/main.go#L345

[spiffxp]

/area kubetest
per the above comment

[spiffxp]

@fejta which method, writeMetadata()?

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

Prevent issues from auto-closing with an /lifecycle frozen comment.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or @fejta.
/lifecycle stale

[spiffxp]

/remove-lifecycle stale
this remains a pain point for consumers of testgrid during release

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[krzyzacy]

/remove-lifecycle stale
/lifecycle frozen

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[spiffxp]

/remove-lifecycle stale
/assign
I wrote a doc about figuring upgrade versions, it would be way nicer to make that doc smaller and say "just look at gubernator"

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[ixdy]

/pony sad

someday someone will work on this.

[k8s-ci-robot]

@ixdy:

In response to this:

/pony sad

someday someone will work on this.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[justaugustus]

/remove-lifecycle rotten
/lifecycle frozen

[cpanato]

@justaugustus i can try this one if i get some initial guide, like where to start :)

[justaugustus]

Thanks Carlos! I personally just found this issue.

@spiffxp @fejta -- Can you provide some details on current state of affairs and what Carlos would need to work on this?

/assign @cpanato

[spiffxp]

This was primarily in reference to all of the jobs that now live here: https://testgrid.k8s.io/sig-release-orphaned

They are run via kubetest: https://github.com/kubernetes/test-infra/tree/master/kubetest#upgrade-skew-kubemark

Testgrid columns are defined in testgrid config.yaml like so: https://github.com/kubernetes/test-infra/blob/master/testgrid/config.md#column-headers

So the questions I can't answer off the top of my head and leave to you are:

• how can you identify the versions being upgraded (is it version files, is it api calls)
• where can you do this (is it a change in kubetest, is it a change in something kubetest kicks off)
• where can you store this data such that it ends up in the metadata section of finished.json (this one could probably use a docs PR when you find the answer)

[cpanato]

@spiffxp Thanks Aaron, since I'm quite new to the codebase I will ping you at slack to talk more, of course, if you agree on that :)

[BenTheElder]

until the rest is agreed there's nothing to do in kubetest, this metadata can just as easily come from the (rather broken) upgrade system.

[spiffxp]

/remove-lifecycle frozen
/assign @BenTheElder
putting this on your radar

[spiffxp]

/remove-priority important-soon

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[cpanato]

this is something we still want to do?
@BenTheElder @spiffxp

[BenTheElder]

IMHO both versions are under test and I think it's debateable which is primary.
I think both versions should be inspected if we're looking at either of them.
@amwat is mostly maintaining kubetest now, in bug-fix mode while we roll out kubetest2.

[spiffxp]

I don't have bandwidth for this, if nobody else does I guess we can let it age out.

This sort of thing remains something the release-team has the power to change if they want, see https://github.com/kubernetes/test-infra/blob/master/testgrid/config.md#column-headers for how to specify what data should appear in what column

It just requires the time to figure out how to get the versions you want, and to make sure they end up in metadata.json (or finished.json's metadata object)

[spiffxp]

I opened #20650 for another testgrid columne hygiene thing I think the community could help with

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-contributor-experience at kubernetes/community.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-contributor-experience at kubernetes/community.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.