寫在前面

上章節(jié)中介紹了Deployment，ReplicaSet，ReplicationController等副本控制器的使用和場景，接下來介紹kubernetes系列教程控制器DaemonSet使用。

1. DaemonSet控制器

1.1 DaemonSet簡介

介紹DaemonSet時(shí)我們先來思考一個(gè)問題：相信大家都接觸過監(jiān)控系統(tǒng)比如zabbix，監(jiān)控系統(tǒng)需要在被監(jiān)控機(jī)安裝一個(gè)agent，安裝agent通常會(huì)涉及到以下幾個(gè)場景：

所有節(jié)點(diǎn)都必須安裝agent以便采集監(jiān)控?cái)?shù)據(jù)
新加入的節(jié)點(diǎn)需要配置agent，手動(dòng)或者運(yùn)行腳本
節(jié)點(diǎn)下線后需要手動(dòng)在監(jiān)控系統(tǒng)中刪除

kubernetes中經(jīng)常涉及到在node上安裝部署應(yīng)用，它是如何解決上述的問題的呢？答案是DaemonSet。DaemonSet守護(hù)進(jìn)程簡稱DS，適用于在所有節(jié)點(diǎn)或部分節(jié)點(diǎn)運(yùn)行一個(gè)daemon守護(hù)進(jìn)程，如監(jiān)控我們安裝部署時(shí)網(wǎng)絡(luò)插件kube-flannel和kube-proxy，DaemonSet具有如下特點(diǎn)：

DaemonSet確保所有節(jié)點(diǎn)運(yùn)行一個(gè)Pod副本
指定節(jié)點(diǎn)運(yùn)行一個(gè)Pod副本，通過標(biāo)簽選擇器或者節(jié)點(diǎn)親和性
新增節(jié)點(diǎn)會(huì)自動(dòng)在節(jié)點(diǎn)增加一個(gè)Pod
移除節(jié)點(diǎn)時(shí)垃圾回收機(jī)制會(huì)自動(dòng)清理Pod

DaemnonSet控制器

DaemonSet適用于每個(gè)node節(jié)點(diǎn)均需要部署一個(gè)守護(hù)進(jìn)程的場景，常見的場景例如：

日志采集agent，如fluentd或logstash
監(jiān)控采集agent，如Prometheus Node Exporter,Sysdig Agent,Ganglia gmond
分布式集群組件，如Ceph MON，Ceph OSD，glusterd，Hadoop Yarn NodeManager等
k8s必要運(yùn)行組件，如網(wǎng)絡(luò)flannel，weave，calico，kube-proxy等

安裝k8s時(shí)默認(rèn)在kube-system命名空間已經(jīng)安裝了有兩個(gè)DaemonSet，分別為kube-flannel-ds-amd64和kube-proxy，分別負(fù)責(zé)flannel overlay網(wǎng)絡(luò)的互通和service代理的實(shí)現(xiàn)，可以通過如下命令查看：

查看kube-system命令空間的DaemonSet列表，當(dāng)前集群有三個(gè)node節(jié)點(diǎn)，所以每個(gè)DS會(huì)運(yùn)行三個(gè)Pod副本

[root@node-1 ~]# kubectl get ds -n kube-system 
NAME                    DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR                   AGE
kube-flannel-ds-amd64   3         3         3       3            3           beta.kubernetes.io/arch=amd64   46d
kube-proxy              3         3         3       3            3           beta.kubernetes.io/os=linux     46d

查看Pod的副本情況，可以看到DaemonSet在每個(gè)節(jié)點(diǎn)都運(yùn)行一個(gè)Pod

image

1.2 DaemonSet定義

DaemonSet的定義和Deployment定義使用相類似，需要定義apiVersion，Kind，metadata和spec屬性信息，spec中不需要定義replicas個(gè)數(shù)，spec.template即定義DS生成容器的模版信息，如下是運(yùn)行一個(gè)fluentd-elasticsearch鏡像容器的daemon守護(hù)進(jìn)程，運(yùn)行在每個(gè)node上通過fluentd采集日志上報(bào)到ElasticSearch。

通過yaml文件定義DaemonSet

[root@node-1 happylau]# cat fluentd-es-daemonset.yaml 
apiVersion: apps/v1              #api版本信息
kind: DaemonSet                  #類型為DaemonSet
metadata:                        #元數(shù)據(jù)信息
  name: fluentd-elasticsearch
  namespace: kube-system        #運(yùn)行的命名空間
  labels:
    k8s-app: fluentd-logging
spec:                          #DS模版
  selector:
    matchLabels:
      name: fluentd-elasticsearch
  template:
    metadata:
      labels:
        name: fluentd-elasticsearch
    spec:
      tolerations:
      - key: node-role.kubernetes.io/master
        effect: NoSchedule
      containers:            #容器信息
      - name: fluentd-elasticsearch
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:          #resource資源
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:      #掛載存儲(chǔ)，agent需要到這些目錄采集日志
        - name: varlog
          mountPath: /var/log
        - name: varlibdockercontainers
          mountPath: /var/lib/docker/containers
          readOnly: true
      terminationGracePeriodSeconds: 30
      volumes:            #將主機(jī)的目錄以hostPath的形式掛載到容器Pod中。
      - name: varlog
        hostPath:
          path: /var/log
      - name: varlibdockercontainers
        hostPath:
          path: /var/lib/docker/containers

DaemonSet定義注意事項(xiàng)：

daemonset.spec.template定義Pod的模板信息，包含的metadata信息需要和selector保持一致
template必須定義RestartPolicy的策略，切策略值為Always，保障服務(wù)異常時(shí)能自動(dòng)重啟恢復(fù)
Pod運(yùn)行在特定節(jié)點(diǎn)，支持指定調(diào)度策略，如nodeSelector，Node affinity，實(shí)現(xiàn)靈活調(diào)度

生成DaemonSet

[root@node-1 happylau]# kubectl apply -f fluentd-es-daemonset.yaml 
daemonset.apps/fluentd-elasticsearch created

查看DaemonSet列表

[root@node-1 happylau]# kubectl get daemonsets -n kube-system  fluentd-elasticsearch 
NAME                    DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
fluentd-elasticsearch   3         3         3       3            3           <none>          16s

查看node上運(yùn)行Pod的情況,在NODE列可以看到每個(gè)node都運(yùn)行了一個(gè)Pod

[root@node-1 happylau]# kubectl get pods -n kube-system -o wide |grep fluentd 
fluentd-elasticsearch-blpqb      1/1     Running   0          3m7s   10.244.2.79      node-3   <none>           <none>
fluentd-elasticsearch-ksdlt      1/1     Running   0          3m7s   10.244.0.11      node-1   <none>           <none>
fluentd-elasticsearch-shtkh      1/1     Running   0          3m7s   10.244.1.64      node-2   <none>           <none>

查看DaemonSet詳情，可以看到DaemonSet支持RollingUpdate滾動(dòng)更新策略

[root@node-1 happylau]# kubectl get daemonsets -n kube-system fluentd-elasticsearch -o yaml
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"apps/v1","kind":"DaemonSet","metadata":{"annotations":{},"labels":{"k8s-app":"fluentd-logging"},"name":"fluentd-elasticsearch","namespace":"kube-system"},"spec":{"selector":{"matchLabels":{"name":"fluentd-elasticsearch"}},"template":{"metadata":{"labels":{"name":"fluentd-elasticsearch"}},"spec":{"containers":[{"image":"quay.io/fluentd_elasticsearch/fluentd:v2.5.2","name":"fluentd-elasticsearch","resources":{"limits":{"memory":"200Mi"},"requests":{"cpu":"100m","memory":"200Mi"}},"volumeMounts":[{"mountPath":"/var/log","name":"varlog"},{"mountPath":"/var/lib/docker/containers","name":"varlibdockercontainers","readOnly":true}]}],"terminationGracePeriodSeconds":30,"tolerations":[{"effect":"NoSchedule","key":"node-role.kubernetes.io/master"}],"volumes":[{"hostPath":{"path":"/var/log"},"name":"varlog"},{"hostPath":{"path":"/var/lib/docker/containers"},"name":"varlibdockercontainers"}]}}}}
  creationTimestamp: "2019-10-30T15:19:20Z"
  generation: 1
  labels:
    k8s-app: fluentd-logging
  name: fluentd-elasticsearch
  namespace: kube-system
  resourceVersion: "6046222"
  selfLink: /apis/extensions/v1beta1/namespaces/kube-system/daemonsets/fluentd-elasticsearch
  uid: c2c02c48-9f93-48f3-9d6c-32bfa671db0e
spec:
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      name: fluentd-elasticsearch
  template:
    metadata:
      creationTimestamp: null
      labels:
        name: fluentd-elasticsearch
    spec:
      containers:
      - image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        imagePullPolicy: IfNotPresent
        name: fluentd-elasticsearch
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        terminationMessagePath: /dev/termination-log
        terminationMessagePolicy: File
        volumeMounts:
        - mountPath: /var/log
          name: varlog
        - mountPath: /var/lib/docker/containers
          name: varlibdockercontainers
          readOnly: true
      dnsPolicy: ClusterFirst
      restartPolicy: Always             #重啟策略必須為Always，保障異常時(shí)能自動(dòng)恢復(fù)
      schedulerName: default-scheduler  #默認(rèn)調(diào)度策略
      securityContext: {}
      terminationGracePeriodSeconds: 30
      tolerations:
      - effect: NoSchedule
        key: node-role.kubernetes.io/master
      volumes:
      - hostPath:
          path: /var/log
          type: ""
        name: varlog
      - hostPath:
          path: /var/lib/docker/containers
          type: ""
        name: varlibdockercontainers
  templateGeneration: 1
  updateStrategy:  #滾動(dòng)更新策略
    rollingUpdate:
      maxUnavailable: 1
    type: RollingUpdate
status:
  currentNumberScheduled: 3
  desiredNumberScheduled: 3
  numberAvailable: 3
  numberMisscheduled: 0
  numberReady: 3
  observedGeneration: 1
  updatedNumberScheduled: 3

1.3 滾動(dòng)更新與回滾

更新鏡像至最新版本

[root@node-1 ~]# kubectl set image daemonsets fluentd-elasticsearch fluentd-elasticsearch=quay.io/fluentd_elasticsearch/fluentd:latest -n kube-system
daemonset.extensions/fluentd-elasticsearch image updated

查看滾動(dòng)更新狀態(tài)

[root@node-1 ~]# kubectl rollout status daemonset -n kube-system fluentd-elasticsearch 
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 1 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 1 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 1 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 2 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 2 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 2 out of 3 new pods have been updated...
Waiting for daemon set "fluentd-elasticsearch" rollout to finish: 2 of 3 updated pods are available...
daemon set "fluentd-elasticsearch" successfully rolled out

查看DaemonSet詳情，可以看到DS滾動(dòng)更新的過程：DaemonSet先將node上的pod刪除然后再創(chuàng)建

DaemonSet滾動(dòng)更新過程

查看DaemonSet滾動(dòng)更新版本，REVSION 1為初始的版本

[root@node-1 ~]# kubectl rollout history daemonset -n kube-system fluentd-elasticsearch 
daemonset.extensions/fluentd-elasticsearch 
REVISION  CHANGE-CAUSE
1         <none>
2         <none>

更新回退，如果配置沒有符合到預(yù)期可以回滾到原始的版本

[root@node-1 ~]# kubectl rollout undo daemonset -n kube-system fluentd-elasticsearch --to-revision=1
daemonset.extensions/fluentd-elasticsearch rolled back

確認(rèn)版本回退情況

DaemonSet版本回退

觀察版本回退的過程，回退的過程和和滾動(dòng)更新過程類似，先刪除Pod再創(chuàng)建

DaemonSet回退過程

刪除DaemonSet

[root@node-1 ~]# kubectl delete daemonsets -n kube-system fluentd-elasticsearch 
daemonset.extensions "fluentd-elasticsearch" deleted
[root@node-1 ~]# kubectl get pods -n kube-system |grep fluentd
fluentd-elasticsearch-d6f6f      0/1     Terminating   0          110m

1.4 DaemonSet調(diào)度

前面kubernetes系列教程（七）深入玩轉(zhuǎn)pod調(diào)度文章介紹了Pod的調(diào)度機(jī)制，DaemonSet通過kubernetes默認(rèn)的調(diào)度器scheduler會(huì)在所有的node節(jié)點(diǎn)上運(yùn)行一個(gè)Pod副本，可以通過如下三種方式將Pod運(yùn)行在部分節(jié)點(diǎn)上：

指定nodeName節(jié)點(diǎn)運(yùn)行
通過標(biāo)簽運(yùn)行nodeSelector
通過親和力調(diào)度node Affinity和node Anti-affinity

DaemonSet調(diào)度算法用于實(shí)現(xiàn)將Pod運(yùn)行在特定的node節(jié)點(diǎn)上，如下以通過node affinity親和力將Pod調(diào)度到部分的節(jié)點(diǎn)上node-2上為例。

為node添加一個(gè)app=web的labels

[root@node-1 happylau]# kubectl get nodes --show-labels 
NAME     STATUS   ROLES    AGE   VERSION   LABELS
node-1   Ready    master   47d   v1.15.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=node-1,kubernetes.io/os=linux,node-role.kubernetes.io/master=
node-2   Ready    <none>   47d   v1.15.3   app=web,beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=node-2,kubernetes.io/os=linux
node-3   Ready    <none>   47d   v1.15.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=node-3,kubernetes.io/os=linux

添加node affinity親和力調(diào)度算法，requiredDuringSchedulingIgnoredDuringExecution設(shè)置基本需要滿足條件，preferredDuringSchedulingIgnoredDuringExecution設(shè)置優(yōu)選滿足條件

[root@node-1 happylau]# cat fluentd-es-daemonset.yaml 
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: fluentd-elasticsearch
  namespace: kube-system
  labels:
    k8s-app: fluentd-logging
spec:
  selector:
    matchLabels:
      name: fluentd-elasticsearch
  template:
    metadata:
      labels:
        name: fluentd-elasticsearch
    spec:
      tolerations:
      - key: node-role.kubernetes.io/master
        effect: NoSchedule
      containers:
      - name: fluentd-elasticsearch
        image: quay.io/fluentd_elasticsearch/fluentd:v2.5.2
        resources:
          limits:
            memory: 200Mi
          requests:
            cpu: 100m
            memory: 200Mi
        volumeMounts:
        - name: varlog
          mountPath: /var/log
        - name: varlibdockercontainers
          mountPath: /var/lib/docker/containers
          readOnly: true
      affinity:
        nodeAffinity:
          preferredDuringSchedulingIgnoredDuringExecution:  #優(yōu)先滿足條件
          - weight: 1
            preference:
              matchExpressions:
              - key: app 
                operator: In
                values:
                - web 
          requiredDuringSchedulingIgnoredDuringExecution:  #要求滿足條件
            nodeSelectorTerms:
            - matchExpressions:
              - key: kubernetes.io/hostname
                operator: In
                values:
                - node-2
                - node-3
      terminationGracePeriodSeconds: 30
      volumes:
      - name: varlog
        hostPath:
          path: /var/log
      - name: varlibdockercontainers
        hostPath:
          path: /var/lib/docker/containers

生成DS，并查看列表

[root@node-1 happylau]# kubectl delete ds -n kube-system fluentd-elasticsearch 
daemonset.extensions "fluentd-elasticsearch" deleted

[root@node-1 happylau]# kubectl get daemonsets -n kube-system fluentd-elasticsearch 
NAME                    DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
fluentd-elasticsearch   1         1         1       1            1           <none>          112s

校驗(yàn)Pod運(yùn)行的情況，DaemonSet的Pod調(diào)度到node-2節(jié)點(diǎn)上

[root@node-1 happylau]# kubectl get pods -n kube-system -o wide 
NAME                             READY   STATUS    RESTARTS   AGE     IP               NODE     NOMINATED NODE   READINESS GATES          <none>
fluentd-elasticsearch-9kngs      1/1     Running   0          2m39s   10.244.1.82      node-2   <none>           <none>

寫在最后

本文介紹了kubernetes中DaemonSet控制器，DS控制器能確保所有的節(jié)點(diǎn)運(yùn)行一個(gè)特定的daemon守護(hù)進(jìn)程，此外通過nodeSelector或node Affinity能夠?qū)崿F(xiàn)將Pod調(diào)度到特定的node節(jié)點(diǎn)。

參考文檔

DaemonSet：https://kubernetes.io/docs/concepts/workloads/controllers/daemonset/

當(dāng)你的才華撐不起你的野心時(shí)，你就應(yīng)該靜下心來學(xué)習(xí)

返回kubernetes系列教程目錄

如果覺得文章對(duì)您有幫助，請訂閱專欄，分享給有需要的朋友吧??

關(guān)于作者劉海平（HappyLau ）云計(jì)算高級(jí)顧問目前在騰訊云從事公有云相關(guān)工作，曾就職于酷狗，EasyStack，擁有多年公有云+私有云計(jì)算架構(gòu)設(shè)計(jì)，運(yùn)維，交付相關(guān)經(jīng)驗(yàn)，參與了酷狗，南方電網(wǎng)，國泰君安等大型私有云平臺(tái)建設(shè)，精通Linux，Kubernetes，OpenStack，Ceph等開源技術(shù)，在云計(jì)算領(lǐng)域具有豐富實(shí)戰(zhàn)經(jīng)驗(yàn)，擁有RHCA/OpenStack/Linux授課經(jīng)驗(yàn)。

色偷偷精品伊人,欧洲久久精品,欧美综合婷婷骚逼,国产AV主播,国产最新探花在线,九色在线视频一区,伊人大交九欧美,1769亚洲,黄色成人av

kubernetes系列教程（十二）詳解DaemonSet控制器

kubernetes系列教程（十二）詳解DaemonSet控制器

寫在前面

1. DaemonSet控制器

1.1 DaemonSet簡介

1.2 DaemonSet定義

1.3 滾動(dòng)更新與回滾

1.4 DaemonSet調(diào)度

寫在最后

參考文檔

相關(guān)閱讀更多精彩內(nèi)容

友情鏈接更多精彩內(nèi)容

色偷偷精品伊人,欧洲久久精品,欧美综合婷婷骚逼,国产AV主播,国产最新探花在线,九色在线视频一区,伊人大交九 欧美,1769亚洲,黄色成人av

kubernetes系列教程（十二）詳解DaemonSet控制器

寫在前面

1. DaemonSet控制器

1.1 DaemonSet簡介

1.2 DaemonSet定義

1.3 滾動(dòng)更新與回滾

1.4 DaemonSet調(diào)度

寫在最后

參考文檔

相關(guān)閱讀更多精彩內(nèi)容

友情鏈接更多精彩內(nèi)容

色偷偷精品伊人,欧洲久久精品,欧美综合婷婷骚逼,国产AV主播,国产最新探花在线,九色在线视频一区,伊人大交九欧美,1769亚洲,黄色成人av