iT邦幫忙

第 12 屆 iThome 鐵人賽

DAY 17
0
DevOps

Docker獸 究極進化 ~~ Kubernetes獸系列 第 17

Day-17 認識Worker Node(Kubernetes)

What is Kubernetes Worker node

Worker node是用來部署容器的地方,也就是運行服務的機器,所以每個Node中必備著能夠建置容器的執行環境,像是Docker等。

Worker node當中主要包含著四個部分,Pod、Kubelet、cAdvisor以及kube-Proxy。在本篇章也會詳細介紹這些重要組件。

https://ithelp.ithome.com.tw/upload/images/20201002/201297374Zxb5oGBE3.png

Node related CLI

Get all nodes

kubectl get nodes

該command能夠取得當下cluster中所有的nodes

$ kubectl get nodes
NAME                                                STATUS   ROLES    AGE     VERSION
gke-my-first-cluster-1-default-pool-dddd2fae-j0k1   Ready    <none>   2d18h   v1.18.6-gke.3504
gke-my-first-cluster-1-default-pool-dddd2fae-rfl8   Ready    <none>   2d18h   v1.18.6-gke.3504
gke-my-first-cluster-1-default-pool-dddd2fae-tz38   Ready    <none>   2d18h   v1.18.6-gke.3504

Describe specific node

kubectl describe nodes <node_name>

該command可以取得指定node的詳細資訊,像是labels、annotation、address...等,並概述幾個node info中的key-value所代表意思

$ kubectl describe nodes gke-my-first-cluster-1-default-pool-dddd2fae-j0k1
Name:               gke-my-first-cluster-1-default-pool-dddd2fae-j0k1
Roles:              <none>
Labels:             beta.kubernetes.io/arch=amd64
                    beta.kubernetes.io/instance-type=g1-small
                    beta.kubernetes.io/os=linux
                    cloud.google.com/gke-nodepool=default-pool
                    cloud.google.com/gke-os-distribution=cos
                    cloud.google.com/machine-family=g1
                    failure-domain.beta.kubernetes.io/region=asia-east1
                    failure-domain.beta.kubernetes.io/zone=asia-east1-a
                    kubernetes.io/arch=amd64
                    kubernetes.io/hostname=gke-my-first-cluster-1-default-pool-dddd2fae-j0k1
                    kubernetes.io/os=linux
                    node.kubernetes.io/instance-type=g1-small
                    topology.kubernetes.io/region=asia-east1
                    topology.kubernetes.io/zone=asia-east1-a
Annotations:        container.googleapis.com/instance_id: 32495144918220960
                    node.alpha.kubernetes.io/ttl: 0
                    node.gke.io/last-applied-node-labels:
                      cloud.google.com/gke-nodepool=default-pool,cloud.google.com/gke-os-distribution=cos,cloud.google.com/machine-family=g1
                    volumes.kubernetes.io/controller-managed-attach-detach: true
CreationTimestamp:  Wed, 23 Sep 2020 20:26:23 +0800
Taints:             <none>
Unschedulable:      false
Conditions:
  Type                          Status  LastHeartbeatTime                 LastTransitionTime                Reason                          Message
  ----                          ------  -----------------                 ------------------                ------                          -------
  CorruptDockerOverlay2         False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   NoCorruptDockerOverlay2         docker overlay2 is functioning properly
  FrequentUnregisterNetDevice   False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   NoFrequentUnregisterNetDevice   node is functioning properly
  FrequentKubeletRestart        False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   NoFrequentKubeletRestart        kubelet is functioning properly
  FrequentDockerRestart         False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   NoFrequentDockerRestart         docker is functioning properly
  FrequentContainerdRestart     False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   NoFrequentContainerdRestart     containerd is functioning properly
  KernelDeadlock                False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   KernelHasNoDeadlock             kernel has no deadlock
  ReadonlyFilesystem            False   Sat, 26 Sep 2020 14:44:01 +0800   Wed, 23 Sep 2020 20:26:24 +0800   FilesystemIsNotReadOnly         Filesystem is not read-only
  NetworkUnavailable            False   Wed, 23 Sep 2020 20:26:24 +0800   Wed, 23 Sep 2020 20:26:24 +0800   RouteCreated                    NodeController create implicit route
  MemoryPressure                False   Sat, 26 Sep 2020 14:44:23 +0800   Wed, 23 Sep 2020 20:26:23 +0800   KubeletHasSufficientMemory      kubelet has sufficient memory available
  DiskPressure                  False   Sat, 26 Sep 2020 14:44:23 +0800   Wed, 23 Sep 2020 20:26:23 +0800   KubeletHasNoDiskPressure        kubelet has no disk pressure
  PIDPressure                   False   Sat, 26 Sep 2020 14:44:23 +0800   Wed, 23 Sep 2020 20:26:23 +0800   KubeletHasSufficientPID         kubelet has sufficient PID available
  Ready                         True    Sat, 26 Sep 2020 14:44:23 +0800   Wed, 23 Sep 2020 20:26:33 +0800   KubeletReady                    kubelet is posting ready status. AppArmor enabled
Addresses:
  InternalIP:   10.140.0.4
  ExternalIP:   35.201.206.74
  InternalDNS:  gke-my-first-cluster-1-default-pool-dddd2fae-j0k1.asia-east1-a.c.oval-compass-290412.internal
  Hostname:     gke-my-first-cluster-1-default-pool-dddd2fae-j0k1.asia-east1-a.c.oval-compass-290412.internal
Capacity:
 attachable-volumes-gce-pd:  15
 cpu:                        1
 ephemeral-storage:          28679952Ki
 hugepages-2Mi:              0
 memory:                     1727548Ki
 pods:                       110
Allocatable:
 attachable-volumes-gce-pd:  15
 cpu:                        940m
 ephemeral-storage:          9251574536
 hugepages-2Mi:              0
 memory:                     1179708Ki
 pods:                       110
System Info:
 Machine ID:                 1b0e9c72c5dbfbab01123c602f399416
 System UUID:                1b0e9c72-c5db-fbab-0112-3c602f399416
 Boot ID:                    f456cdc3-20d3-406a-9368-738b39a74b19
 Kernel Version:             5.4.49+
 OS Image:                   Container-Optimized OS from Google
 Operating System:           linux
 Architecture:               amd64
 Container Runtime Version:  docker://19.3.9
 Kubelet Version:            v1.18.6-gke.3504
 Kube-Proxy Version:         v1.18.6-gke.3504
PodCIDR:                     10.0.2.0/24
PodCIDRs:                    10.0.2.0/24
ProviderID:                  gce://oval-compass-290412/asia-east1-a/gke-my-first-cluster-1-default-pool-dddd2fae-j0k1
Non-terminated Pods:         (4 in total)
  Namespace                  Name                                                            CPU Requests  CPU Limits  Memory Requests  Memory Limits  AGE
  ---------                  ----                                                            ------------  ----------  ---------------  -------------  ---
  kube-system                konnectivity-agent-kjnbk                                        0 (0%)        0 (0%)      30Mi (2%)        30Mi (2%)      2d18h
  kube-system                kube-dns-56b5cc7c4c-zzxl6                                       260m (27%)    0 (0%)      110Mi (9%)       210Mi (18%)    2d18h
  kube-system                kube-proxy-gke-my-first-cluster-1-default-pool-dddd2fae-j0k1    100m (10%)    0 (0%)      0 (0%)           0 (0%)         2d18h
  kubernetes-dashboard       kubernetes-dashboard-7b544877d5-tth5k                           0 (0%)        0 (0%)      0 (0%)           0 (0%)         25m
Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  Resource                   Requests     Limits
  --------                   --------     ------
  cpu                        360m (38%)   0 (0%)
  memory                     140Mi (12%)  240Mi (20%)
  ephemeral-storage          0 (0%)       0 (0%)
  attachable-volumes-gce-pd  0            0
Events:
  Type     Reason            Age                   From                                                               Message
  ----     ------            ----                  ----                                                               -------
  Warning  NodeSysctlChange  22m (x12 over 2d18h)  sysctl-monitor, gke-my-first-cluster-1-default-pool-dddd2fae-j0k1

Tips: 這些key值與labels都會因為公有雲不同而有所差異。

Addresses

  • HostName: 該節點的host name,可以透過kubectl —hostname-override 來覆寫他。
  • ExternalIP: 該節點可路由的外部IP,提供群集外部使用。
  • InternalIP: 該節點可路由的內部IP,僅叢集內部能夠路由。

Conditions

描述所有運行節點目前的狀態,狀態的描述有以下幾種:

Node Condition Description
Ready True表示節點運行狀況良好並準備好接受Pod,False表示節點運行狀況不佳並且不接受Pod,Unknown表示節點控制器最近一次未從節點收到消息node-monitor-grace-period(默認值為40秒)
DiskPressure True表示磁盤容量不足;除此以外False
MemoryPressure True表示節點內存不足; 除此以外False
PIDPressure True表示節點上的Process太多;除此以外False
NetworkUnavailable True表示節點的網絡配置不正確,否則 False

Capacity and Allocatable

描述該節點上可用資源最大數量,包含cpu、memory與pods的數量等...。

System Info

該節點上各種軟硬體設備的訊息,包含uuid與版本號....等

Pod

Pod是在kubernetes當中,能夠創建與運行的最小執行單位,在Pod當中能夠有著一個或多個Containers,並且這些Containers共享著Pod的資源。因為Pod在Kubernetes當中是相當重要的一個component,因此會在下個篇章詳細介紹!

Kubelet

Kubernetes是一個分散式的集群管理系統,在每個worker 上運行一個worker process對node上的Container做周期性管理,而這個worker就是Kubelet。

kubelet有著以下幾個主要功能

  • Pod的管理: 如上述,一個pod由一或多個containers組成,彼此共享pod中的資源與port,所以同個pod間能透過localhost進行溝通,因此也可利用volume與mount將資源共享至多個容器當中,kubelet就是負責管理這些pod資源
  • 健康檢查: 創建容器後,如果想確認容器是否正常啟動,可以加入health check在pod/deployment的yaml當中,再啟動pod時kubelet會去執行yaml中的health check,只要health check沒過kubelet就會刪除該pod並依照重啟策略處理(預設為刪除後不進行重啟)
  • 容器監測: 透過建置cAdvisor進行監測。

CAdvisor

cAdvisor是一個worker,並即時性的對該Node上所有的資源與容器進行監測與數據的採集,像是CPU、Memory的用量、網路的流量與Storage的使用量等。cAdvisor集成於Kubelet當中,當使用Kubelet時會自動地啟動cAdvisor。

Kube-Proxy

在kubernetes中,網路是相當重要的一環,那簡單來說每個pod都會有個ip,但pod是經常在發生變化的,每次更新ip位置都會有變。為此kubernetes有個component叫做service,每個service都會有一組固定的虛擬ip(clusterIp),並且自動地綁定某種類型的pod,有點類似某種pod的專用通道,所有對於該類型pod的request都會透過service進行load balance與redirect。為了實現該功能,在每個Node上都會有個Kube-Proxy,得以當作service, api-server與pod間溝通的橋樑。

在未來service的篇章會再與service一起詳述。

https://ithelp.ithome.com.tw/upload/images/20201002/20129737lEny7hhLuh.png

小結

在本篇章我們瀏覽的Worker Node的整體架構與初步了解了其中的components,並且初步認識幾個重要的Kubernetes components像是Pod與Service,那在後面的篇章將會為大家延伸並深入認識這些重要components,敬請期待。

https://ithelp.ithome.com.tw/upload/images/20201002/20129737HNTHXKvKVp.png

Reference

https://kubernetes.io/docs/concepts/architecture/nodes/


上一篇
Day-16 認識 Master Node (Kubernetes)
下一篇
Day-18 學習Pod, ReplicaSet 與 Deployment (上)
系列文
Docker獸 究極進化 ~~ Kubernetes獸30

尚未有邦友留言

立即登入留言