kubeadm: remove ClusterConfiguration.DNS.Type from v1beta3

[neolit123]

What this PR does / why we need it:

Kubeadm no longer supports kube-dns and CoreDNS is the only
supported DNS server. Remove ClusterConfiguration.DNS.Type
from v1beta3 that is used to set the DNS server type.

Which issue(s) this PR fixes:

Fixes kubernetes/kubeadm#2398
xref kubernetes/kubeadm#1796

Special notes for your reviewer:

Does this PR introduce a user-facing change?

kubeadm: remove the field ClusterConfiguration.DNS.Type in v1beta3 since CoreDNS is the only supported DNS type.

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

[neolit123]

/kind deprecation cleanup
/priority important-soon
/triage accepted
/cc @fabriziopandini @SataQiu @pacoxu

[neolit123]

/hold
for review

[pacoxu]

/retest

[pacoxu]

/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: neolit123, pacoxu

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:

• cmd/kubeadm/OWNERS [neolit123]

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

[pacoxu]

As node-local-dns is GAed since 1.18, is there a plan to add node-local-dns to ClusterConfiguration.DNS.Cache=true in kubeadm?

[neolit123]

@pacoxu

As node-local-dns is GAed since 1.18, is there a plan to add node-local-dns to ClusterConfiguration.DNS.Cache=true in kubeadm?

we had a few discussions about this in the past.
kubernetes/kubeadm#1220

as an addon, my main argument against it is that its not something that is exactly required for a cluster to work.
kubeadm tries to be minimal and only apply essentials and security defaults. the node local cache would be something that is like a "nice to have", but not essential, and today users can apply it on demand.

we also want to move CoreDNS and similar addons outside of kubeadm to operators or a different form of addon plugins.

[neolit123]

@fabriziopandini i have updated the PR. looking for LGTM.

getting this right is difficult and other fuzzer magic did not help.

the scenario around DNS.Type unravels some limitations in the fuzzers + stock API machinery:

• field exists in the old public API
• field is removed in the latest public API
• field must exists in the internal type, since validation is performed on internal types only and until the old public API is removed.

i argue that this model is flawed and it forces API maintainers to have their latest public API converters and internal APIs know about outdated fields (see cmd/kubeadm/app/apis/kubeadm/v1beta3/conversion.go).

if you are wondering, during fuzzing the Schema converters do execute the defaulting for v1beta2, which means DNS.Type receives CoreDNS if the value is "".

[fabriziopandini]

/lgtm

We can eventually iterate on the fuzzer/internal types on separated PRs given that this isn't user facing.
IMO, what is really important in the long term, is that we find a way to untangle defaulting & the fuzzing logic

[neolit123]

/hold cancel
/test pull-kubernetes-node-e2e-containerd

[fejta-bot]

/retest
This bot automatically retries jobs that failed/flaked on approved PRs (send feedback to fejta).

Review the full test history for this PR.

Silence the bot with an /lgtm cancel or /hold comment for consistent failures.

[fejta-bot]

/retest
This bot automatically retries jobs that failed/flaked on approved PRs (send feedback to fejta).

Review the full test history for this PR.

Silence the bot with an /lgtm cancel or /hold comment for consistent failures.

[fejta-bot]

/retest
This bot automatically retries jobs that failed/flaked on approved PRs (send feedback to fejta).

Review the full test history for this PR.

Silence the bot with an /lgtm cancel or /hold comment for consistent failures.

[fejta-bot]

/retest
This bot automatically retries jobs that failed/flaked on approved PRs (send feedback to fejta).

Review the full test history for this PR.

Silence the bot with an /lgtm cancel or /hold comment for consistent failures.