The documentation is misleading that kube-proxy and kubelet are not on control-planes nodes

engineater commented 3 months ago

The documentation is misleading that kube-proxy and kubelet are not on control-planes nodes.

On page https://kubernetes.io/docs/concepts/architecture/ you can see image that do not show components kube-proxy and kubelet on control-plane node. I think it is not one mistake in the documetation.

k8s_arch_2024-07-08

What would you like to be added On page https://kubernetes.io/docs/concepts/architecture/ in the diagram in "CONTROL PLANE" block need add kube-proxy and kubelet blocks. And I think need to add "CRI" block in "CONTROL PLANE" block too.

k8s-ci-robot commented 3 months ago

This issue is currently awaiting triage.

SIG Docs takes a lead on issue triage for this website, but any Kubernetes member can accept issues by applying the triage/accepted label.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available [here](https://git.k8s.io/community/contributors/guide/pull-requests.md). If you have questions or suggestions related to my behavior, please file an issue against the [kubernetes-sigs/prow](https://github.com/kubernetes-sigs/prow/issues/new?title=Prow%20issue:) repository.

tamilselvan1102 commented 3 months ago

page related to the issue: https://kubernetes.io/docs/concepts/architecture/

tamilselvan1102 commented 3 months ago

Kubernetes Worker Nodes only contain kubelet, kube-proxy

tamilselvan1102 commented 3 months ago

Each node contains a kubelet, which communicate with the Kubernetes control plane. All nodes contain kube-proxy, that facilitates Kubernetes networking services.

tamilselvan1102 commented 3 months ago

There is no issue with the diagram. it depicts correctly

engineater commented 3 months ago

Control-plane is different role of node, but when I make technical interview about k8s. a candidates 100% sure that kubelet, kube-proxy are not on control-planes. I think the image is the problem.

tamilselvan1102 commented 3 months ago

yes, it is correct kubelet, kube-proxy are not on control-planes. The image also shows same.

engineater commented 3 months ago

I use k8s 1.29.5 deploy with standard kubeadm. So I can see that control-plane has kube-proxy.

$ kubectl get no -o wide
NAME                        STATUS   ROLES            AGE   VERSION   INTERNAL-IP     EXTERNAL-IP   OS-IMAGE                  KERNEL-VERSION                     CONTAINER-RUNTIME
ol9-master-151.my.private   Ready    control-plane    42d   v1.29.5   192.168.1.151   <none>        Oracle Linux Server 9.4   5.15.0-206.153.7.1.el9uek.x86_64   cri-o://1.29.4
ol9-worker-154.my.private   Ready    ingress,worker   42d   v1.29.5   192.168.1.154   <none>        Oracle Linux Server 9.4   5.15.0-206.153.7.1.el9uek.x86_64   cri-o://1.29.4
ol9-worker-155.my.private   Ready    ingress,worker   42d   v1.29.5   192.168.1.155   <none>        Oracle Linux Server 9.4   5.15.0-206.153.7.1.el9uek.x86_64   containerd://1.6.32

$ kubectl get po -A -o wide | grep proxy
kube-system          kube-proxy-59p9n                                    1/1     Running   16              37d   192.168.1.154    ol9-worker-154.my.private   <none>           <none>
kube-system          kube-proxy-9q7sz                                    1/1     Running   16              37d   192.168.1.151    ol9-master-151.my.private   <none>           <none>
kube-system          kube-proxy-j2vcf                                    1/1     Running   17 (144m ago)   37d   192.168.1.155    ol9-worker-155.my.private   <none>           <none>

$ kubectl get po kube-proxy-9q7sz -n kube-system -o yaml
apiVersion: v1
kind: Pod
metadata:
  creationTimestamp: "2024-06-01T04:49:49Z"
  generateName: kube-proxy-
  labels:
    controller-revision-hash: f966846b6
    k8s-app: kube-proxy
    pod-template-generation: "2"
  name: kube-proxy-9q7sz
  namespace: kube-system
  ownerReferences:
  - apiVersion: apps/v1
    blockOwnerDeletion: true
    controller: true
    kind: DaemonSet
    name: kube-proxy
    uid: 3721bef7-4306-43ff-85fb-e474916e4668
  resourceVersion: "168239"
  uid: 4b84e26a-cc63-443d-a260-1194073c001f
spec:
. . . . .
status:
  conditions:
  - lastProbeTime: null
    lastTransitionTime: "2024-07-08T07:25:04Z"
    status: "True"
    type: PodReadyToStartContainers
  - lastProbeTime: null
    lastTransitionTime: "2024-06-01T04:49:49Z"
    status: "True"
    type: Initialized
  - lastProbeTime: null
    lastTransitionTime: "2024-07-08T07:25:04Z"
    status: "True"
    type: Ready
  - lastProbeTime: null
    lastTransitionTime: "2024-07-08T07:25:04Z"
    status: "True"
    type: ContainersReady
  - lastProbeTime: null
    lastTransitionTime: "2024-06-01T04:49:49Z"
    status: "True"
    type: PodScheduled
  containerStatuses:
  - containerID: cri-o://3bcbf5e11da200df6b191d8b00c3d7f389934a97f9d24c956c2ad7a949c324c7
    image: registry.k8s.io/kube-proxy:v1.29.5
    imageID: registry.k8s.io/kube-proxy@sha256:4c9681a68b0f068f66e6c4120be71a4416621cad1427802deaaa79d01fdffb85
    lastState: {}
    name: kube-proxy
    ready: true
    restartCount: 16
    started: true
    state:
      running:
        startedAt: "2024-07-08T07:25:04Z"
  hostIP: 192.168.1.151
  hostIPs:
  - ip: 192.168.1.151
  phase: Running
  podIP: 192.168.1.151
  podIPs:
  - ip: 192.168.1.151
  qosClass: BestEffort
  startTime: "2024-06-01T04:49:49Z"

engineater commented 3 months ago

On control-plane I can see that kubelet is running

$ sudo pstree |  grep kubelet
        |-kubelet---15*[{kubelet}]

$ systemctl status kubelet.service
● kubelet.service - kubelet: The Kubernetes Node Agent
     Loaded: loaded (/usr/lib/systemd/system/kubelet.service; enabled; preset: disabled)
    Drop-In: /usr/lib/systemd/system/kubelet.service.d
             └─10-kubeadm.conf
     Active: active (running) since Mon 2024-07-08 12:24:37 +05; 2h 45min ago
       Docs: https://kubernetes.io/docs/
   Main PID: 1396 (kubelet)
      Tasks: 16 (limit: 12185)
     Memory: 74.2M
        CPU: 3min 7.576s
     CGroup: /system.slice/kubelet.service
             └─1396 /usr/bin/kubelet --bootstrap-kubeconfig=/etc/kubernetes/bootstrap-kubelet.conf --kubeconfig=/etc/kubernetes/kubelet.conf --config=/var/lib/kubelet/config.yaml --container-runtime-endpoint=unix:///var/run/crio/crio.sock --pod-infra-container-image=registry.k8s.io/pause:3.9
. . . .

kundan2707 commented 3 months ago

@engineater control-plane and master node both are different things. all components running on master nodes= are not control-place component. kube-proxy and kubelete run on every master as well as worker node. control-place component are those gives instruction to other nodes and manage cluster.

engineater commented 3 months ago

@kundan2707 well in k8s 1.29 (maby start from 1.27) we do not have concept "master" node. We have only concepts "control-plane" and "workers" nodes. So if we say "control plane" that is mean "control-plane". No "master" node, because we do not have it.

For my prove

$ kubectl get no ol9-master-151.my.private -o yaml | grep master
    csi.volume.kubernetes.io/nodeid: '{"csi.tigera.io":"ol9-master-151.my.private"}'
    kubernetes.io/hostname: ol9-master-151.my.private
  name: ol9-master-151.my.private
  - address: ol9-master-151.my.private

$ kubectl get no ol9-master-151.my.private -o yaml | grep control
    volumes.kubernetes.io/controller-managed-attach-detach: "true"
    node-role.kubernetes.io/control-plane: ""
    key: node-role.kubernetes.io/control-plane
    - registry.k8s.io/kube-controller-manager@sha256:a9a64e67b66ea6fb43f976f65d8a0cadd68b0ed5ed2311d2fc4bf887403ecf8a
    - registry.k8s.io/kube-controller-manager@sha256:acdb952db121aa1e8182d70b36ceb868020e0435b7b8fd016dda1346acbc22a3
    - registry.k8s.io/kube-controller-manager:v1.29.5

engineater commented 3 months ago

Well if the diagram show architecture whitout mistakes. Is it possible on page https://kubernetes.io/docs/concepts/architecture/ add some description that kubelet and kube-proxy also run on control-plane nodes? O maby add some couple phrases?

tengqm commented 3 months ago

control-plane is equivalent to "master" and it refers to a collection of critical components that manage the essential of the cluster concept. control-plane nodes are nodes that are meant to run the control-plane components. There are different setups in production. For example, the control-plane components can be running on any of the "cluster nodes". They may be managed by the vendor rather than the cluster user. They can be deployed as static Pods or systemd services. When they are deployed as static Pods, you will need "kubelet" to bring them up, otherwise, you may and may not need kubelet at all.

There are many setups where the control-plane nodes are also used to deploy workload pods. You may want to deploy cluster management components such as authentication, 3rd-party webhook servers, monitoring, logging services there. If your control-plane nodes are big enough, you may want to host pods from some trusted namespaces as well.

In all, the whole thing is a pretty flexible. You choose your own topology, based on your business requirement, rather than what the typical setup is. They may and may not suit your needs.

engineater commented 3 months ago

@tengqm I think you explanation will be good to see on document https://kubernetes.io/docs/concepts/architecture/ .

neolit123 commented 3 months ago

the easiest fix is for the linked page to mention that this is an example reference architecture. additionally someone could expand with a paragraph of text what @tengqm has shared in his comment: https://github.com/kubernetes/website/issues/47111#issuecomment-2214009124

/help

k8s-ci-robot commented 3 months ago

@neolit123: This request has been marked as needing help from a contributor.

Guidelines

Please ensure that the issue body includes answers to the following questions:

Why are we solving this issue?
To address this issue, are there any code changes? If there are code changes, what needs to be done in the code and what places can the assignee treat as reference points?
Does this issue have zero to low barrier of entry?
How can the assignee reach out to you for help?

For more details on the requirements of such an issue, please see here and ensure that they are met.

If this request no longer meets these requirements, the label can be removed by commenting with the /remove-help command.

In response to [this](https://github.com/kubernetes/website/issues/47111): >the easiest fix is for the linked page to mention that this is an example reference architecture. >additionally someone could expand with a paragraph of text what @tengqm has shared in his comment: >https://github.com/kubernetes/website/issues/47111#issuecomment-2214009124 > >/help Instructions for interacting with me using PR comments are available [here](https://git.k8s.io/community/contributors/guide/pull-requests.md). If you have questions or suggestions related to my behavior, please file an issue against the [kubernetes-sigs/prow](https://github.com/kubernetes-sigs/prow/issues/new?title=Prow%20issue:) repository.

neolit123 commented 3 months ago

/sig architecture

robert-cronin commented 2 months ago

I have some ideas for this issue if no one has started already! /assign

neolit123 commented 2 months ago

thanks for the PR @robert-cronin . commented https://github.com/kubernetes/website/pull/47164

sftim commented 2 months ago

@engineater would #47164 help here?

kubernetes / website

The documentation is misleading that kube-proxy and kubelet are not on control-planes nodes #47111

Guidelines