Migrate to new networking.k8s.io/v1beta1 package

[aledbf]

What this PR does / why we need it:

Migrates from the extensions package to the new one networking.k8s.io/v1beta1.
There are some restrictions:

• We cannot break existing installations in clusters < 1.14
• This is the reason why e2e tests keep the old package
• Internally we detect the version and convert the object to the new one
• After k8s 1.21 (three releases after 1.18 we can change the tests and remove the extensions package)

Migrates from the "k8s.io/api/networking/v1beta1" package to "k8s.io/api/networking/v1beta1" for Kubernetes versions >= 1.14.0 and keeps the extension package for older versions.

Which issue this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close that issue when PR gets merged): fixes #

replaces #4081

Special notes for your reviewer:

[cmluciano]

@aledbf Will this replace #4081 ?

[aledbf]

@cmluciano not sure yet

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@aa131b0). Click here to learn what that means.
The diff coverage is 62.69%.

@@            Coverage Diff            @@
##             master    #4127   +/-   ##
=========================================
  Coverage          ?   57.48%           
=========================================
  Files             ?       87           
  Lines             ?     6534           
  Branches          ?        0           
=========================================
  Hits              ?     3756           
  Misses            ?     2345           
  Partials          ?      433

Impacted Files	Coverage Δ
internal/ingress/types.go	0% <ø> (ø)
internal/ingress/controller/store/ingress.go	0% <0%> (ø)
cmd/nginx/main.go	6.57% <0%> (ø)
internal/k8s/main.go	60.37% <0%> (ø)
internal/ingress/annotations/ratelimit/main.go	61.68% <100%> (ø)
internal/ingress/annotations/influxdb/main.go	60% <100%> (ø)
internal/ingress/annotations/cors/main.go	57.14% <100%> (ø)
internal/ingress/annotations/ipwhitelist/main.go	55.17% <100%> (ø)
internal/ingress/annotations/snippet/main.go	100% <100%> (ø)
internal/ingress/annotations/luarestywaf/main.go	60% <100%> (ø)
... and 37 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update aa131b0...5f8deb2. Read the comment docs.

[aledbf]

@cmluciano ok, pull-ingress-nginx-test job is now passing and only two e2e tests are failing in pull-ingress-nginx-e2e-1-14.

[cmluciano]

Cool, do you think those leader elect failures are related to the api change moving between different versions?

[aledbf]

Cool, do you think those leader elect failures are related to the api change moving between different versions?

Yes. This error is not present in other passing e2e tests

No. This was related to missing permission to change the configmap when the ingress class is testclass

[aledbf]

/test pull-ingress-nginx-test

[aledbf]

/test all

[aledbf]

/retest

[aledbf]

/test pull-ingress-nginx-e2e-1-12

[aledbf]

/retest

[ElvinEfendi]

/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: aledbf, ElvinEfendi

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 [ElvinEfendi,aledbf]

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

[nealAR]

We have non-RBAC enabled cluster running 1.14.6 on AKS. After upgrading NGINX to 0.25.0 our ingress will no longer resolve to anything but the default route (all other routes currently return a 404). The release notes for 0.25.0 state "please make sure to update the ClusterRole required for the new networking.k8s.io/v1beta1 Ingress API", so I'm thinking the issue may have something do to with this PR? A couple question: 1) does that assumption sound correction given the issues we're having with our ingress? 2) is an update to the ClusterRole required to go along with this PR? 3) if either 1 or 2 are correct how would we resolve the issue for a non-RBAC enabled cluster?

[nealAR]

Also are we required to change the apiVersion of the ingress-controller from networking.k8s.io/v1beta1 to networking/v1beta1?

[davidkarlsen]

@aledbf
Should the release still run on on k8s 1.13.9 ? I experience that the ingress resources do not get assigned LB address with the upgraded nginx-ingress. Using metallb for providing LoadBalancer IPs.