Refactor cgroup hierarchy handling and resource reservation

* We don't need to organize the cgroup hierarchy differently if we don't
  use the resources reservation, so remove the variance, always place
  the kubelet at the same place (default to
  /runtime.slice/kubelet.service)

* Same for the container "runtimes" (which means in fact the container
  **engines**, aka containerd, cri-o, not runc or kata)

* Accordingly, there is no need for a lot of customization on the cgroup
  hierarchy, so reduce it to `kube_slice` and `system_slice`. All the
  rest is derived from that and not user-modifiable.

* Correct the semantics of kube_reserved and system_reserved:
  - kube-reserved and systemd-reserved do not guarantee on their own
    that resources will be available for the respective cgroups, they
    allow to calculate NodeAllocatable.
    See https://kubernetes.io/docs/tasks/administer-cluster/reserve-compute-resources/#node-allocatable

roles/container-engine/containerd/templates/containerd.service.j2

@@ -36,10 +36,8 @@ LimitMEMLOCK={{ containerd_limit_mem_lock }}
 # Only systemd 226 and above support this version.
 TasksMax=infinity
 OOMScoreAdjust=-999
-# Set the cgroup slice of the service so that kube reserved takes effect
-{% if kube_reserved is defined and kube_reserved|bool %}
-Slice={{ kube_reserved_cgroups_for_service_slice }}
-{% endif %}
+# Set the cgroup slice of the service to optionally enforce resource limitations
+Slice={{ kube_slice }}
 
 [Install]
 WantedBy=multi-user.target

roles/container-engine/cri-dockerd/templates/cri-dockerd.service.j2

@@ -35,10 +35,8 @@ LimitCORE=infinity
 TasksMax=infinity
 Delegate=yes
 KillMode=process
-# Set the cgroup slice of the service so that kube reserved takes effect
-{% if kube_reserved is defined and kube_reserved|bool %}
-Slice={{ kube_reserved_cgroups_for_service_slice }}
-{% endif %}
+# Set the cgroup slice of the service to optionally enforce resource limitations
+Slice={{ kube_slice }}
 
 [Install]
 WantedBy=multi-user.target

roles/container-engine/cri-o/tasks/main.yaml

@@ -90,19 +90,16 @@
     remote_src: true
   notify: Restart crio
 
-- name: Cri-o | configure crio to use kube reserved cgroups
+- name: Cri-o | configure crio to run in the kube slice
   ansible.builtin.copy:
     dest: /etc/systemd/system/crio.service.d/00-slice.conf
     owner: root
     group: root
     mode: '0644'
     content: |
       [Service]
-      Slice={{ kube_reserved_cgroups_for_service_slice }}
+      Slice={{ kube_slice }}
   notify: Restart crio
-  when:
-    - kube_reserved is defined and kube_reserved is true
-    - kube_reserved_cgroups_for_service_slice is defined
 
 - name: Cri-o | update the bin dir for crio.service file
   replace:

roles/container-engine/cri-o/templates/crio.conf.j2

@@ -114,11 +114,7 @@ conmon = "{{ crio_conmon }}"
 {% if crio_cgroup_manager == "cgroupfs" %}
 conmon_cgroup = "pod"
 {% else %}
-{% if kube_reserved is defined and kube_reserved|bool %}
-conmon_cgroup = "{{ kube_reserved_cgroups_for_service_slice }}"
-{% else %}
-conmon_cgroup = "system.slice"
-{% endif %}
+conmon_cgroup = "{{ kube_slice }}"
 {% endif %}
 
 # Environment variable list for the conmon process, used for passing necessary

roles/container-engine/docker/templates/docker.service.j2

@@ -32,10 +32,8 @@ TimeoutStartSec=1min
 Restart=on-failure
 StartLimitBurst=3
 StartLimitInterval=60s
-# Set the cgroup slice of the service so that kube reserved takes effect
-{% if kube_reserved is defined and kube_reserved|bool %}
-Slice={{ kube_reserved_cgroups_for_service_slice }}
-{% endif %}
+# Set the cgroup slice of the service to optionally enforce resource limitations
+Slice={{ kube_slice }}
 
 [Install]
 WantedBy=multi-user.target

roles/kubernetes/node/defaults/main.yml

@@ -11,15 +11,6 @@ kube_resolv_conf: "/etc/resolv.conf"
 # Set to empty to avoid cgroup creation
 kubelet_enforce_node_allocatable: "\"\""
 
-# Set runtime and kubelet cgroups when using systemd as cgroup driver (default)
-kube_service_cgroups: "{% if kube_reserved %}{{ kube_reserved_cgroups_for_service_slice }}{% else %}system.slice{% endif %}"
-kubelet_runtime_cgroups: "/{{ kube_service_cgroups }}/{{ container_manager }}.service"
-kubelet_kubelet_cgroups: "/{{ kube_service_cgroups }}/kubelet.service"
-
-# Set runtime and kubelet cgroups when using cgroupfs as cgroup driver
-kubelet_runtime_cgroups_cgroupfs: "/system.slice/{{ container_manager }}.service"
-kubelet_kubelet_cgroups_cgroupfs: "/system.slice/kubelet.service"
-
 # Set systemd service hardening features
 kubelet_systemd_hardening: false
 
@@ -42,6 +33,10 @@ kube_cpu_reserved: "100m"
 kube_ephemeral_storage_reserved: "500Mi"
 kube_pid_reserved: "1000"
 
+# Set slice for host system daemons (sshd, NetworkManager, ...)
+# You probably don't want to change this
+system_slice: system.slice
+
 # Set to true to reserve resources for system daemons
 system_reserved: false
 system_reserved_cgroups_for_service_slice: system.slice

roles/kubernetes/node/tasks/facts.yml

@@ -39,12 +39,6 @@
     kubelet_cgroup_driver: "{{ kubelet_cgroup_driver_detected }}"
   when: kubelet_cgroup_driver is undefined
 
-- name: Set kubelet_cgroups options when cgroupfs is used
-  set_fact:
-    kubelet_runtime_cgroups: "{{ kubelet_runtime_cgroups_cgroupfs }}"
-    kubelet_kubelet_cgroups: "{{ kubelet_kubelet_cgroups_cgroupfs }}"
-  when: kubelet_cgroup_driver == 'cgroupfs'
-
 - name: Set kubelet_config_extra_args options when cgroupfs is used
   set_fact:
     kubelet_config_extra_args: "{{ kubelet_config_extra_args | combine(kubelet_config_extra_args_cgroupfs) }}"

roles/kubernetes/node/templates/kubelet-config.v1beta1.yaml.j2

@@ -33,7 +33,7 @@ address: {{ kubelet_bind_address }}
 readOnlyPort: {{ kube_read_only_port }}
 healthzPort: {{ kubelet_healthz_port }}
 healthzBindAddress: {{ kubelet_healthz_bind_address }}
-kubeletCgroups: {{ kubelet_kubelet_cgroups }}
+kubeletCgroups: {{ kube_slice_cgroup ~ 'kubelet.service' }}
 clusterDomain: {{ dns_domain }}
 {% if kubelet_protect_kernel_defaults | bool %}
 protectKernelDefaults: true
@@ -63,7 +63,7 @@ clusterDNS:
 {# Node reserved CPU/memory #}
 {% for scope in "kube", "system" %}
 {% if lookup('ansible.builtin.vars', scope + "_reserved") | bool %}
-{{ scope }}ReservedCgroup: {{ lookup('ansible.builtin.vars', scope + '_reserved_cgroups') }}
+{{ scope }}ReservedCgroup: {{ lookup('ansible.builtin.vars', scope + '_slice_cgroup') }}
 {% endif %}
 {{ scope }}Reserved:
 {% for resource in "cpu", "memory", "ephemeral-storage", "pid" %}

roles/kubernetes/node/templates/kubelet.env.v1beta1.j2

@@ -11,7 +11,7 @@ KUBELET_HOSTNAME="--hostname-override={{ kube_override_hostname }}"
 --config={{ kube_config_dir }}/kubelet-config.yaml \
 --kubeconfig={{ kube_config_dir }}/kubelet.conf \
 {# end kubeadm specific settings #}
---runtime-cgroups={{ kubelet_runtime_cgroups }} \
+--runtime-cgroups={{ kube_slice_cgroup ~ container_manager ~ '.service' }} \
 {% endset %}
 
 KUBELET_ARGS="{{ kubelet_args_base }} {{ kubelet_custom_flags | join(' ') }}"

roles/kubernetes/node/templates/kubelet.service.j2

@@ -14,9 +14,7 @@ Wants={{ kubelet_dependency }}
 {% endfor %}
 
 [Service]
-{% if kube_reserved|bool %}
-Slice={{ kube_reserved_cgroups_for_service_slice }}
-{% endif %}
+Slice={{ kube_slice }}
 EnvironmentFile=-{{ kube_config_dir }}/kubelet.env
 ExecStart={{ bin_dir }}/kubelet \
 		$KUBE_LOGTOSTDERR \

roles/kubernetes/node/vars/main.yml

@@ -0,0 +1,3 @@
+---
+kube_slice_cgroup: "/{{ kube_slice.split('-') | join('.slice/') }}/"
+system_slice_cgroup: "/{{ system_slice.split('-') | join('.slice/') }}/"

roles/kubespray-defaults/defaults/main/main.yml

@@ -23,6 +23,11 @@ kube_version: v1.30.4
 ## The minimum version working
 kube_version_min_required: v1.28.0
 
+# TODO: put this default to more specific place -> needed by roles container-engine+kubernetes/node
+# Set the systemd slice for kubernetes-related daemons: kubelet and container engine
+# You probably don't want to change this
+kube_slice: runtime.slice
+
 ## Kube Proxy mode One of ['iptables', 'ipvs']
 kube_proxy_mode: ipvs