Refactor control plane upgrades with reconfiguration support (#12015)

* Refactor control plane upgrades with reconfiguration support

Adds revised support for:
- The previously removed `--config` argument for `kubeadm upgrade apply`
- Changes to `ClusterConfiguration` as part of the `upgrade-cluster.yml` playbook lifecycle
- kubeadm-config `v1beta4` `UpgradeConfiguration` for the `kubeadm upgrade apply` command: [UpgradeConfiguration v1beta4](https://kubernetes.io/docs/reference/config-api/kubeadm-config.v1beta4/#kubeadm-k8s-io-v1beta4-UpgradeConfiguration).

* Add kubeadm upgrade node support

Per discussion:
- Use `kubeadm upgrade node` on secondary control plane upgrades
- Add support for UpgradeConfiguration.node in kubeadm-config.v1beta4
- Remove redundant `allowRCUpgrades` config
- Revert from `block` for first and secondary control plane back to unblocked tasks since they no longer share much code and it's more readable this way

* Add kubelet and kube-proxy reconfiguration to upgrades

* Fix task to use `kubeadm init phase etcd local`

* Rebase with changes from "Adapt checksums and versions to new hashes updater" PR

* Add `imagePullPolicy` and `imagePullSerial` to kubeadm-config v1beta4 `InitConfiguration.nodeRegistration`

roles/kubernetes/control-plane/defaults/main/main.yml

@@ -248,6 +248,9 @@ kube_apiserver_tracing_sampling_rate_per_million: 100
 # Enable kubeadm file discovery if anonymous access has been removed
 kubeadm_use_file_discovery: "{{ remove_anonymous_access }}"
 
+# imagePullSerial specifies if image pulling performed by kubeadm must be done serially or in parallel. Default: true
+kubeadm_image_pull_serial: true
+
 # Supported asymmetric encryption algorithm types for the cluster's keys and certificates.
 # can be one of RSA-2048(default), RSA-3072, RSA-4096, ECDSA-P256
 # ref: https://kubernetes.io/docs/reference/config-api/kubeadm-config.v1beta4/#kubeadm-k8s-io-v1beta4-ClusterConfiguration

roles/kubernetes/control-plane/tasks/check-api.yml

@@ -0,0 +1,10 @@
+---
+- name: Kubeadm | Check api is up
+  uri:
+    url: "https://{{ ip | default(fallback_ip) }}:{{ kube_apiserver_port }}/healthz"
+    validate_certs: false
+  when: ('kube_control_plane' in group_names)
+  register: _result
+  retries: 60
+  delay: 5
+  until: _result.status == 200

roles/kubernetes/control-plane/tasks/kubeadm-setup.yml

@@ -229,7 +229,7 @@
 - name: Kubeadm | Join other control plane nodes
   include_tasks: kubeadm-secondary.yml
 
-- name: Kubeadm | upgrade kubernetes cluster
+- name: Kubeadm | upgrade kubernetes cluster to {{ kube_version }}
   include_tasks: kubeadm-upgrade.yml
   when:
     - upgrade_cluster_setup

roles/kubernetes/control-plane/tasks/kubeadm-upgrade.yml

@@ -1,56 +1,81 @@
 ---
-- name: Kubeadm | Check api is up
-  uri:
-    url: "https://{{ main_ip | ansible.utils.ipwrap }}:{{ kube_apiserver_port }}/healthz"
-    validate_certs: false
-  when: ('kube_control_plane' in group_names)
-  register: _result
-  retries: 60
-  delay: 5
-  until: _result.status == 200
+- name: Ensure kube-apiserver is up before upgrade
+  import_tasks: check-api.yml
 
+  # kubeadm-config.v1beta4 with UpgradeConfiguration requires some values that were previously allowed as args to be specified in the config file
 - name: Kubeadm | Upgrade first control plane node
   command: >-
     timeout -k 600s 600s
-    {{ bin_dir }}/kubeadm
-    upgrade apply -y v{{ kube_version }}
+    {{ bin_dir }}/kubeadm upgrade apply -y v{{ kube_version }}
+    {%- if kubeadm_config_api_version == 'v1beta3' %}
     --certificate-renewal={{ kubeadm_upgrade_auto_cert_renewal }}
     --ignore-preflight-errors={{ kubeadm_ignore_preflight_errors | join(',') }}
     --allow-experimental-upgrades
     --etcd-upgrade={{ (etcd_deployment_type == "kubeadm") | lower }}
     {% if kubeadm_patches | length > 0 %}--patches={{ kubeadm_patches_dir }}{% endif %}
     --force
+    {%- else %}
+    --config={{ kube_config_dir }}/kubeadm-config.yaml
+    {%- endif -%}
   register: kubeadm_upgrade
-  # Retry is because upload config sometimes fails
-  retries: 3
-  until: kubeadm_upgrade.rc == 0
   when: inventory_hostname == first_kube_control_plane
   failed_when: kubeadm_upgrade.rc != 0 and "field is immutable" not in kubeadm_upgrade.stderr
   environment:
     PATH: "{{ bin_dir }}:{{ ansible_env.PATH }}"
-  notify: Control plane | restart kubelet
 
 - name: Kubeadm | Upgrade other control plane nodes
   command: >-
-    timeout -k 600s 600s
-    {{ bin_dir }}/kubeadm
-    upgrade apply -y v{{ kube_version }}
+    {{ bin_dir }}/kubeadm upgrade node
+    {%- if kubeadm_config_api_version == 'v1beta3' %}
     --certificate-renewal={{ kubeadm_upgrade_auto_cert_renewal }}
     --ignore-preflight-errors={{ kubeadm_ignore_preflight_errors | join(',') }}
-    --allow-experimental-upgrades
     --etcd-upgrade={{ (etcd_deployment_type == "kubeadm") | lower }}
     {% if kubeadm_patches | length > 0 %}--patches={{ kubeadm_patches_dir }}{% endif %}
-    --force
+    {%- else %}
+    --config={{ kube_config_dir }}/kubeadm-config.yaml
+    {%- endif -%}
   register: kubeadm_upgrade
-  # Retry is because upload config sometimes fails
-  retries: 3
-  until: kubeadm_upgrade.rc == 0
   when: inventory_hostname != first_kube_control_plane
   failed_when: kubeadm_upgrade.rc != 0 and "field is immutable" not in kubeadm_upgrade.stderr
   environment:
     PATH: "{{ bin_dir }}:{{ ansible_env.PATH }}"
+
+# kubeadm upgrade no longer reconciles ClusterConfiguration and KubeProxyConfiguration changes, this must be done separately after upgrade to ensure the latest config is applied
+- name: Update kubeadm and kubelet configmaps after upgrade
+  command: "{{ bin_dir }}/kubeadm init phase upload-config all --config {{ kube_config_dir }}/kubeadm-config.yaml"
+  register: kubeadm_upload_config
+  # Retry is because upload config sometimes fails
+  retries: 3
+  until: kubeadm_upload_config.rc == 0
+  when:
+  - inventory_hostname == first_kube_control_plane
+
+- name: Update kube-proxy configmap after upgrade
+  command: "{{ bin_dir }}/kubeadm init phase addon kube-proxy --config {{ kube_config_dir }}/kubeadm-config.yaml"
+  register: kube_proxy_upload_config
+  # Retry is because upload config sometimes fails
+  retries: 3
+  until: kube_proxy_upload_config.rc == 0
+  when:
+  - inventory_hostname == first_kube_control_plane
+  - ('addon/kube-proxy' not in kubeadm_init_phases_skip)
+
+- name: Rewrite kubeadm managed etcd static pod manifests with updated configmap
+  command: "{{ bin_dir }}/kubeadm init phase etcd local --config {{ kube_config_dir }}/kubeadm-config.yaml"
+  when:
+  - etcd_deployment_type == "kubeadm"
   notify: Control plane | restart kubelet
 
+- name: Rewrite kubernetes control plane static pod manifests with updated configmap
+  command: "{{ bin_dir }}/kubeadm init phase control-plane all --config {{ kube_config_dir }}/kubeadm-config.yaml"
+  notify: Control plane | restart kubelet
+
+- name: Flush kubelet handlers
+  meta: flush_handlers
+
+- name: Ensure kube-apiserver is up after upgrade and control plane configuration updates
+  import_tasks: check-api.yml
+
 - name: Kubeadm | Remove binding to anonymous user
   command: "{{ kubectl }} -n kube-public delete rolebinding kubeadm:bootstrap-signer-clusterinfo --ignore-not-found"
   when: remove_anonymous_access
@@ -60,8 +85,8 @@
     path: "{{ item }}"
     state: absent
   with_items:
-    - /root/.kube/cache
-    - /root/.kube/http-cache
+  - /root/.kube/cache
+  - /root/.kube/http-cache
 
 # FIXME: https://github.com/kubernetes/kubeadm/issues/1318
 - name: Kubeadm | scale down coredns replicas to 0 if not using coredns dns_mode
@@ -75,6 +100,6 @@
   until: scale_down_coredns is succeeded
   run_once: true
   when:
-    - kubeadm_scale_down_coredns_enabled
-    - dns_mode not in ['coredns', 'coredns_dual']
+  - kubeadm_scale_down_coredns_enabled
+  - dns_mode not in ['coredns', 'coredns_dual']
   changed_when: false

roles/kubernetes/control-plane/templates/kubeadm-config.v1beta4.yaml.j2

@@ -29,6 +29,8 @@ nodeRegistration:
   - name: cloud-provider
     value: external
 {% endif %}
+  imagePullPolicy: {{ k8s_image_pull_policy }}
+  imagePullSerial: {{ kubeadm_image_pull_serial | lower }}
 {% if kubeadm_patches | length > 0 %}
 patches:
   directory: {{ kubeadm_patches_dir }}
@@ -480,6 +482,42 @@ scheduler:
 {% endfor %}
 {% endif %}
 ---
+apiVersion: kubeadm.k8s.io/v1beta4
+kind: UpgradeConfiguration
+apply:
+  kubernetesVersion: v{{ kube_version }}
+  allowExperimentalUpgrades: true
+  certificateRenewal: {{ kubeadm_upgrade_auto_cert_renewal | lower }}
+  etcdUpgrade: {{ (etcd_deployment_type == "kubeadm") | lower }}
+  forceUpgrade: true
+{% if kubeadm_ignore_preflight_errors | length > 0 %}
+  ignorePreflightErrors:
+{% for ignore_error in kubeadm_ignore_preflight_errors %}
+  - "{{ ignore_error }}"
+{% endfor %}
+{% endif %}
+{% if kubeadm_patches | length > 0 %}
+  patches:
+    directory: {{ kubeadm_patches_dir }}
+{% endif %}
+  imagePullPolicy: {{ k8s_image_pull_policy }}
+  imagePullSerial: {{ kubeadm_image_pull_serial | lower }}
+node:
+  certificateRenewal: {{ kubeadm_upgrade_auto_cert_renewal | lower }}
+  etcdUpgrade: {{ (etcd_deployment_type == "kubeadm") | lower }}
+{% if kubeadm_ignore_preflight_errors | length > 0 %}
+  ignorePreflightErrors:
+{% for ignore_error in kubeadm_ignore_preflight_errors %}
+  - "{{ ignore_error }}"
+{% endfor %}
+{% endif %}
+{% if kubeadm_patches | length > 0 %}
+  patches:
+    directory: {{ kubeadm_patches_dir }}
+{% endif %}
+  imagePullPolicy: {{ k8s_image_pull_policy }}
+  imagePullSerial: {{ kubeadm_image_pull_serial | lower }}
+---
 apiVersion: kubeproxy.config.k8s.io/v1alpha1
 kind: KubeProxyConfiguration
 bindAddress: "{{ kube_proxy_bind_address }}"