Bump CAPI to v1.5.0

[mboersma]

What type of PR is this?

/kind cleanup

What this PR does / why we need it:

Updates CAPI to v1.5.0 and all that entails.

Which issue(s) this PR fixes:

Fixes #3680
Closes #3751
Closes #3773

Special notes for your reviewer:

• cherry-pick candidate

TODOs:

• squashed commits
• includes documentation
• adds unit tests

Release note:

Bump CAPI to v1.5.0

[mboersma]

/retitle [WIP] Bump CAPI to v1.5.0-rc.0

[mboersma]

I have a handle on fixing the remaining errors related to controller-runtime, except for this one:

# github.com/Azure/azure-service-operator/v2/api/resources/v1api20200601
../../../pkg/mod/github.com/!azure/azure-service-operator/v2@v2.1.0/api/resources/v1api20200601/resource_group_types_gen.go:178:29: cannot use &ResourceGroup{} (value of type *ResourceGroup) as admission.Validator value in variable declaration: *ResourceGroup does not implement admission.Validator (wrong type for method ValidateCreate)
		have ValidateCreate() error
		want ValidateCreate() (admission.Warnings, error)
make: *** [Makefile:198: manager] Error 1

@nojnhuh does this imply we need a new release of ASO?

[nojnhuh]

I have a handle on fixing the remaining errors related to controller-runtime, except for this one:

# github.com/Azure/azure-service-operator/v2/api/resources/v1api20200601
../../../pkg/mod/github.com/!azure/azure-service-operator/v2@v2.1.0/api/resources/v1api20200601/resource_group_types_gen.go:178:29: cannot use &ResourceGroup{} (value of type *ResourceGroup) as admission.Validator value in variable declaration: *ResourceGroup does not implement admission.Validator (wrong type for method ValidateCreate)
		have ValidateCreate() error
		want ValidateCreate() (admission.Warnings, error)
make: *** [Makefile:198: manager] Error 1

@nojnhuh does this imply we need a new release of ASO?

Yes, that's what it looks like to me. ASO is currently on controller-runtime 0.13:
https://github.com/Azure/azure-service-operator/blob/7131774f3f7eb7cc65e431a30eafcb1ab363921f/v2/go.mod#L42

[nojnhuh]

A temporary alternative to waiting for an ASO release would be ripping out all of our ASO resource group stuff.

[mboersma]

I opened a request for controller-runtime v0.15.0 support at Azure/azure-service-operator#3137. Hopefully they're ok with upgrading–it is a breaking change.

[Jont828]

@mboersma Yeah that's unfortunate that the versions of controller runtime don't line up. Is there anything we can do for that besides wait for them to release a new version after upgrading to v0.15.0?

[Jont828]

LGTM pending the ASO issue

[mboersma]

Is there anything we can do for that besides wait...?

We could back out the current ASO implementation for resource groups if need be, but that wouldn't really change anything, just buy some more time.

Maybe we could make an adaptor-type struct that would override ValidateCreate and the other methods and just return an empty admission.Warnings to satisfy the new func signatures. I've been looking around though and it's difficult to see where that code would go.

[Jont828]

Sounds good. Do you have an idea of when we want to get this PR in? We would need to get this in order to merge #3736 as it requires bumping to v1.5 as well.

[mboersma]

when we want to get this PR in?

As soon as we can after the CAPI release, as always (planned for July 25), but we will also need to get a new ASO release and we don't control that timeline. The good news is that team pounced on our request and already have a PR in review that upgrades controller-runtime to match CAPI 1.5, so hopefully an ASO release isn't far away.

[nojnhuh]

@mboersma Here are the changes to include ASO v2.2.0 which includes controller-runtime 0.15: 28b88459

It looks like ASO's CRD manifest and docker image for the new release aren't out yet, but I've asked about that in Slack and did a quick smoke test with locally-built versions and I expect everything to work once things are published.

[nojnhuh]

It looks like ASO's CRD manifest and docker image for the new release aren't out yet, but I've asked about that in Slack and did a quick smoke test with locally-built versions and I expect everything to work once things are published.

ASO release kinks appear to have been worked out.

[mboersma]

/retest

[mboersma]

@Jont828 AFAICT it's ready to go, but we've had several flakes in the main e2e test, including the one I noted above. That test hasn't been that reliable, and the flakes seem to be similar to ones on other PRs.

My only concern is that perhaps v1.5.0 is introducing some further instability there, but it's really hard to discern. I haven't been able to reproduce any of these failures, nor track down any specifics through the logs. They don't seem to be related to this PR, but I can't say so for sure.

[Jont828]

Sounds good, I'm comfortable moving this forward if we're seeing similar flakes in other PRs.

[mboersma]

Rebased due to a collision in test code from #3826.

[CecileRobertMichon]

I looked at the e2e test history and I don't see a pattern in the failures

[mboersma]

Rebased again due to collision in go.sum with #3827.

[willie-yao]

/lgtm

[k8s-ci-robot]

LGTM label has been added.

Git tree hash: 93fa01bf8fc44fc51c13bc358dfa977ef4f6e75b

[mboersma]

/test pull-cluster-api-provider-azure-e2e
/test pull-cluster-api-provider-azure-e2e-optional

[mboersma]

/test pull-cluster-api-provider-azure-e2e-optional

[CecileRobertMichon]

this is really cool! we should use those, there are a few places where I'm sure they'd be useful

[CecileRobertMichon]

/lgtm
/approve
/hold

feel free to remove hold when optional tests have passed

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: CecileRobertMichon

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [CecileRobertMichon]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[mboersma]

/hold cancel

[mboersma]

/override pull-cluster-api-provider-azure-apidiff

[k8s-ci-robot]

@mboersma: Overrode contexts on behalf of mboersma: pull-cluster-api-provider-azure-apidiff

In response to this:

/override pull-cluster-api-provider-azure-apidiff

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.

[Jont828]

We made it!