kubernetes核心概念 Controller

一、pod控制器controller

1.1 Controller作用及分類

controller用于控制pod

參考: https://kubernetes.io/zh/docs/concepts/workloads/controllers/

控制器主要分為:

• Deployments 部署無(wú)狀態(tài)應(yīng)用，控制pod升級(jí),回退
• ReplicaSet 副本集,控制pod擴(kuò)容,裁減
• ReplicationController(相當(dāng)于ReplicaSet的老版本,現(xiàn)在建議使用Deployments加ReplicaSet替代RC)
• StatefulSets 部署有狀態(tài)應(yīng)用，結(jié)合Service、存儲(chǔ)等實(shí)現(xiàn)對(duì)有狀態(tài)應(yīng)用部署
• DaemonSet 守護(hù)進(jìn)程集，運(yùn)行在所有集群節(jié)點(diǎn)(包括master), 比如使用filebeat,node_exporter
• Jobs 一次性
• Cronjob 周期性

1.2 Deployment

1.2.1 Replicaset控制器的功能

• 支持新的基于集合的selector(以前的rc里沒(méi)有這種功能)
• 通過(guò)改變Pod副本數(shù)量實(shí)現(xiàn)Pod的擴(kuò)容和縮容

1.2.2 Deployment控制器的功能

• Deployment集成了上線部署、滾動(dòng)升級(jí)、創(chuàng)建副本、回滾等功能
• Deployment里包含并使用了ReplicaSet

1.2.3 Deployment用于部署無(wú)狀態(tài)應(yīng)用

無(wú)狀態(tài)應(yīng)用的特點(diǎn):

• 所有pod無(wú)差別
• 所有pod中容器運(yùn)行同一個(gè)image
• 所有pod可以運(yùn)行在集群中任意node上
• 所有pod無(wú)啟動(dòng)順序先后之分
• 隨意pod數(shù)量擴(kuò)容或縮容
• 例如簡(jiǎn)單運(yùn)行一個(gè)靜態(tài)web程序

1.2.4 創(chuàng)建deployment類型應(yīng)用

1, 準(zhǔn)備YAML文件

[root@k8s-master1 ~]# vim deployment-nginx.yml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: deploy-nginx			# deployment名
spec:				
  replicas: 1					# 副本集,deployment里使用了replicaset
  selector:
    matchLabels:
      app: nginx				# 匹配的pod標(biāo)簽,表示deployment和rs控制器控制帶有此標(biāo)簽的pod
  template:					    # 代表pod的配置模板
    metadata:
      labels:
        app: nginx				# pod的標(biāo)簽
    spec:
      containers:				# 以下為pod里的容器定義
      - name: nginx
        image: nginx:1.15-alpine
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80

2, 應(yīng)用YAML文件創(chuàng)建deployment

[root@k8s-master1 ~]# kubectl apply -f deployment-nginx.yml
deployment.apps/deploy-nginx created

3, 查看驗(yàn)證

[root@k8s-master1 ~]# kubectl get deployment				# deployment可簡(jiǎn)寫成depoly
NAME           READY   UP-TO-DATE   AVAILABLE   AGE
deploy-nginx   1/1     1            1           19s

[root@k8s-master1 ~]# kubectl get pods
NAME                            READY   STATUS    RESTARTS   AGE
deploy-nginx-6c9764bb69-pbc2h   1/1     Running   0          75s

[root@k8s-master1 ~]# kubectl get replicasets				# replicasets可簡(jiǎn)寫成rs
NAME                      DESIRED   CURRENT   READY   AGE
deploy-nginx-6c9764bb69   1         1         1       2m6s

1.2.5 訪問(wèn)deployment

1,查看pod的IP地址

[root@k8s-master1 ~]# kubectl get pods -o wide
NAME                            READY   STATUS    RESTARTS   AGE   IP               NODE          NOMINATED NODE   READINESS GATES
deploy-nginx-6d9d558bb6-88nr8   1/1     Running   0          39s   10.244.159.155   k8s-master1   <none>           <none>

pod在k8s-master1節(jié)點(diǎn),pod的IP為10.244.159.155

2, 查看所有集群節(jié)點(diǎn)的網(wǎng)卡

[root@k8s-master1 ~]# ifconfig tunl0 |head -2
tunl0: flags=193<UP,RUNNING,NOARP>  mtu 1480
        inet 10.244.159.128  netmask 255.255.255.255

[root@k8s-master2 ~]# ifconfig tunl0 |head -2
tunl0: flags=193<UP,RUNNING,NOARP>  mtu 1480
        inet 10.244.224.0  netmask 255.255.255.255

[root@k8s-master3 ~]# ifconfig tunl0 |head -2
tunl0: flags=193<UP,RUNNING,NOARP>  mtu 1480
        inet 10.244.135.192  netmask 255.255.255.255

[root@k8s-worker1 ~]# ifconfig tunl0 |head -2
tunl0: flags=193<UP,RUNNING,NOARP>  mtu 1480
        inet 10.244.194.64  netmask 255.255.255.255

• 可以看到所有集群節(jié)點(diǎn)的IP都為10.244.0.0/16這個(gè)大網(wǎng)段內(nèi)的子網(wǎng)

3, 在任意集群節(jié)點(diǎn)上都可以訪問(wèn)此deploy里pod

# curl 10.244.159.155

結(jié)果是任意集群節(jié)點(diǎn)都可以訪問(wèn)這個(gè)POD,但集群外部是不能訪問(wèn)的

1.2.6 刪除deployment中的pod

1, 刪除pod（注意: 是刪除deployment中的pod）

[root@k8s-master1 ~]# kubectl delete pod deploy-nginx-6c9764bb69-pbc2h
pod "deploy-nginx-6c9764bb69-pbc2h" deleted

2, 再次查看,發(fā)現(xiàn)又重新啟動(dòng)了一個(gè)pod(節(jié)點(diǎn)由k8s-master1轉(zhuǎn)為k8s-worker1 了,IP地址也變化了)

[root@k8s-master1 ~]# kubectl get pods  -o wide
NAME                            READY   STATUS    RESTARTS   AGE   IP              NODE          NOMINATED NODE   READINESS GATES
deploy-nginx-6d9d558bb6-f2t6r   1/1     Running   0          28s   10.244.194.94   k8s-worker1   <none>           <none>

也就是說(shuō)**pod的IP不是固定的,比如把整個(gè)集群關(guān)閉再啟動(dòng),pod也會(huì)自動(dòng)啟動(dòng),但是IP地址也會(huì)變化**

既然IP地址不是固定的,所以需要一個(gè)固定的訪問(wèn)endpoint給用戶,那么這種方式就是service.

1.2.7 pod版本升級(jí)

查看幫助

[root@k8s-master1 ~]# kubectl set image -h

1, 升級(jí)前驗(yàn)證nginx版本

[root@k8s-master1 ~]# kubectl describe pods deploy-nginx-6d9d558bb6-f2t6r | grep Image:
    Image:          nginx:1.15-alpine
    
[root@k8s-master1 ~]# kubectl exec deploy-nginx-6d9d558bb6-f2t6r -- nginx -v
nginx version: nginx/1.15.12

2, 升級(jí)為1.16版

[root@k8s-master1 ~]# kubectl set image deployment deploy-nginx nginx=nginx:1.16-alpine --record
deployment.apps/deploy-nginx image updated

說(shuō)明:

• deployment deploy-nginx代表名為deploy-nginx的deployment

• nginx=nginx:1.16-alpine前面的nginx為容器名

• –record 表示會(huì)記錄

容器名怎么查看?

• kubectl describe pod pod名查看
• kubectl edit deployment deployment名來(lái)查看容器名
• kubectl get deployment deployment名 -o yaml來(lái)查看容器名

3, 驗(yàn)證

如果升級(jí)的pod數(shù)量較多，則需要一定時(shí)間，可通過(guò)下面命令查看是否已經(jīng)成功

[root@k8s-master1 ~]# kubectl rollout status deployment deploy-nginx	
deployment "deploy-nginx" successfully rolled out

驗(yàn)證 pod

[root@k8s-master1 ~]# kubectl get pods
NAME                            READY   STATUS    RESTARTS   AGE
deploy-nginx-5f4749c8c8-nskp9   1/1     Running   0          104s     更新后,后面的id變了

驗(yàn)證版本

[root@k8s-master1 ~]# kubectl describe pod deploy-nginx-5f4749c8c8-nskp9 |grep Image:
    Image:          nginx:1.16-alpine							升級(jí)為1.16了
    
[root@k8s-master1 ~]# kubectl exec deploy-nginx-5f4749c8c8-nskp9 -- nginx -v
nginx version: nginx/1.16.1										升級(jí)為1.16了

練習(xí): 再將nginx1升級(jí)為1.17版

[root@k8s-master1 ~]# kubectl set image deployment deploy-nginx nginx=nginx:1.17-alpine --record
deployment.apps/deploy-nginx image updated

1.2.8 pod版本回退

1, 查看版本歷史信息

[root@k8s-master1 ~]# kubectl rollout history deployment deploy-nginx
deployment.apps/deploy-nginx
REVISION  CHANGE-CAUSE
1         <none>												原1.15版
2         kubectl set image deployment deploy-nginx nginx=nginx:1.16-alpine --record=true
3         kubectl set image deployment deploy-nginx nginx=nginx:1.17-alpine --record=true

2, 定義要回退的版本（還需要執(zhí)行才是真的回退版本)

[root@k8s-master1 ~]# kubectl rollout history deployment deploy-nginx --revision=1
deployment.apps/deploy-nginx with revision #1
Pod Template:
  Labels:       app=nginx
        pod-template-hash=6c9764bb69
  Containers:
   nginx:
    Image:      nginx:1.15-alpine				可以看到這是要回退的1.15版本
    Port:       80/TCP
    Host Port:  0/TCP
    Environment:        <none>
    Mounts:     <none>
  Volumes:      <none>

3, 執(zhí)行回退

[root@k8s-master1 ~]# kubectl rollout undo deployment deploy-nginx --to-revision=1
deployment.apps/deploy-nginx rolled back

4, 驗(yàn)證

[root@k8s-master1 ~]# kubectl rollout history deployment deploy-nginx
deployment.apps/deploy-nginx
REVISION  CHANGE-CAUSE
2         kubectl set image deployment deploy-nginx nginx=nginx:1.16-alpine --record=true
3         kubectl set image deployment deploy-nginx nginx=nginx:1.17-alpine --record=true
4         <none>						回到了1.15版,但revision的ID變了

[root@k8s-master1 ~]# kubectl get pods
NAME                            READY   STATUS    RESTARTS   AGE
deploy-nginx-6c9764bb69-zgwpj   1/1     Running   0          54s

[root@k8s-master1 ~]# kubectl describe pod deploy-nginx-6c9764bb69-zgwpj |grep Image:
    Image:          nginx:1.15-alpine				回到了1.15版

[root@k8s-master1 ~]# kubectl exec deploy-nginx-6c9764bb69-zgwpj -- nginx -v
nginx version: nginx/1.15.12						回到了1.15版

1.2.9 副本擴(kuò)容

查看幫助

[root@k8s-master1 ~]# kubectl scale -h

1, 擴(kuò)容為2個(gè)副本

[root@k8s-master1 ~]# kubectl scale deployment deploy-nginx --replicas=2
deployment.apps/deploy-nginx scaled

2, 查看

[root@k8s-master1 ~]# kubectl get pods -o wide
NAME                            READY   STATUS    RESTARTS   AGE   IP               NODE          NOMINATED NODE   READINESS GATES
deploy-nginx-6d9d558bb6-4c64l   1/1     Running   0          27s   10.244.159.157   k8s-master1   <none>           <none>
deploy-nginx-6d9d558bb6-hkq2b   1/1     Running   0          71s   10.244.194.95    k8s-worker1   <none>           <none>

在兩個(gè)node節(jié)點(diǎn)上各1個(gè)pod

3, 繼續(xù)擴(kuò)容(我們這里只有2個(gè)node,但是可以大于node節(jié)點(diǎn)數(shù)據(jù))

[root@master ~]# kubectl scale deployment deploy-nginx --replicas=4
deployment.extensions/nginx1 scaled

[root@k8s-master1 ~]# kubectl get pods -o wide
NAME                            READY   STATUS    RESTARTS   AGE     IP               NODE          NOMINATED NODE   READINESS GATES
deploy-nginx-6d9d558bb6-4c64l   1/1     Running   0          87s     10.244.159.157   k8s-master1   <none>           <none>
deploy-nginx-6d9d558bb6-586dr   1/1     Running   0          31s     10.244.135.197   k8s-master3   <none>           <none>
deploy-nginx-6d9d558bb6-hkq2b   1/1     Running   0          2m11s   10.244.194.95    k8s-worker1   <none>           <none>
deploy-nginx-6d9d558bb6-kvgsc   1/1     Running   0          31s     10.244.224.13    k8s-master2   <none>           <none>

1.2.10 副本裁減

1, 指定副本數(shù)為1進(jìn)行裁減

[root@k8s-master1 ~]# kubectl scale deployment deploy-nginx --replicas=1
deployment.apps/deploy-nginx scaled

2, 查看驗(yàn)證

[root@k8s-master1 ~]# kubectl get pods
NAME                            READY   STATUS    RESTARTS   AGE
deploy-nginx-6d9d558bb6-hkq2b   1/1     Running   0          2m56s

1.2.11 多副本滾動(dòng)更新

1, 先擴(kuò)容多點(diǎn)副本

[root@k8s-master1 ~]# kubectl scale deployment deploy-nginx --replicas=16
deployment.apps/deploy-nginx scaled

2, 驗(yàn)證

[root@master ~]# kubectl get pods
NAME                      READY   STATUS    RESTARTS   AGE
nginx1-7d9b8757cf-2hd48   1/1     Running   0          61s
nginx1-7d9b8757cf-5m72n   1/1     Running   0          61s
nginx1-7d9b8757cf-5w2xr   1/1     Running   0          61s
nginx1-7d9b8757cf-5wmdh   1/1     Running   0          61s
nginx1-7d9b8757cf-6szjj   1/1     Running   0          61s
nginx1-7d9b8757cf-9dgsw   1/1     Running   0          61s
nginx1-7d9b8757cf-dc7qj   1/1     Running   0          61s
nginx1-7d9b8757cf-l52pr   1/1     Running   0          61s
nginx1-7d9b8757cf-m7rt4   1/1     Running   0          26m
nginx1-7d9b8757cf-mdkj2   1/1     Running   0          61s
nginx1-7d9b8757cf-s79kp   1/1     Running   0          61s
nginx1-7d9b8757cf-shhvk   1/1     Running   0          61s
nginx1-7d9b8757cf-sv8gb   1/1     Running   0          61s
nginx1-7d9b8757cf-xbhf4   1/1     Running   0          61s
nginx1-7d9b8757cf-zgdgd   1/1     Running   0          61s
nginx1-7d9b8757cf-zzljl   1/1     Running   0          61s
nginx2-559567f789-8hstz   1/1     Running   1          114m

3, 滾動(dòng)更新

[root@k8s-master1 ~]# kubectl set image deployment deploy-nginx nginx=nginx:1.17-alpine --record
deployment.apps/deploy-nginx image updated

4, 驗(yàn)證

[root@k8s-master1 ~]# kubectl rollout status deployment deploy-nginx
......
Waiting for deployment "deploy-nginx" rollout to finish: 13 of 16 updated replicas are available...
Waiting for deployment "deploy-nginx" rollout to finish: 14 of 16 updated replicas are available...
Waiting for deployment "deploy-nginx" rollout to finish: 15 of 16 updated replicas are available...
deployment "deploy-nginx" successfully rolled out

1.2.12 刪除deployment

如果使用 kubectl delete deployment deploy-nginx 命令刪除deployment,那么里面的pod也會(huì)被自動(dòng)刪除

1.3 Replicaset

1, 編寫YAML文件

[root@master ~]# vim rs-nginx.yml
apiVersion: apps/v1
kind: ReplicaSet
metadata:
  name: rs-nginx
  namespace: default
spec:                    # replicaset的spec
  replicas: 2            # 副本數(shù)
  selector:              # 標(biāo)簽選擇器,對(duì)應(yīng)pod的標(biāo)簽
    matchLabels:
      app: nginx         # 匹配的label
  template:
    metadata:
      name: nginx		# pod名
      labels:           # 對(duì)應(yīng)上面定義的標(biāo)簽選擇器selector里面的內(nèi)容
        app: nginx
    spec:               # pod的spec
      containers:
      - name: nginx
        image: nginx:1.15-alpine
        ports:
        - name: http
          containerPort: 80

2, 應(yīng)用YAML文件

[root@k8s-master1 ~]# kubectl apply -f rs-nginx.yml
replicaset.apps/rs-nginx created

3, 驗(yàn)證

[root@k8s-master1 ~]# kubectl get rs
NAME       DESIRED   CURRENT   READY   AGE
rs-nginx   2         2         2       26s

[root@k8s-master1 ~]# kubectl get pods
NAME             READY   STATUS    RESTARTS   AGE
rs-nginx-7j9hz   1/1     Running   0          44s
rs-nginx-pncsk   1/1     Running   0          43s

[root@k8s-master1 ~]# kubectl get deployment
No resources found.

找不到deployment,說(shuō)明創(chuàng)建rs并沒(méi)有創(chuàng)建deployment

二、pod控制器Controller進(jìn)階

2.1 DaemonSet

2.1.1 DaemonSet介紹

• DaemonSet能夠讓所有（或者特定）的節(jié)點(diǎn)運(yùn)行同一個(gè)pod。
• 當(dāng)節(jié)點(diǎn)加入到K8S集群中，pod會(huì)被（DaemonSet）調(diào)度到該節(jié)點(diǎn)上運(yùn)行，當(dāng)節(jié)點(diǎn)從K8S集群中被移除，被DaemonSet調(diào)度的pod會(huì)被移除
• 如果刪除DaemonSet，所有跟這個(gè)DaemonSet相關(guān)的pods都會(huì)被刪除。
• 如果一個(gè)DaemonSet的Pod被殺死、停止、或者崩潰，那么DaemonSet將會(huì)重新創(chuàng)建一個(gè)新的副本在這臺(tái)計(jì)算節(jié)點(diǎn)上。
• DaemonSet一般應(yīng)用于日志收集、監(jiān)控采集、分布式存儲(chǔ)守護(hù)進(jìn)程等

2.1.2 DaemonSet應(yīng)用案例

1, 編寫YAML文件

[root@master ~]# vim daemonset-nginx.yml
apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: daemonset-nginx			
spec:
  selector:
    matchLabels:
      name: nginx-ds
  template:
    metadata:
      labels:
        name: nginx-ds
    spec:
      tolerations:						# tolerations代表容忍
      - key: node-role.kubernetes.io/master  # 能容忍的污點(diǎn)key
        effect: NoSchedule   # kubectl explain pod.spec.tolerations查看(能容忍的污點(diǎn)effect)
      containers:
      - name: nginx
        image: nginx:1.15-alpine
        imagePullPolicy: IfNotPresent
        resources:    # resources資源限制是為了防止master節(jié)點(diǎn)的資源被占太多(根據(jù)實(shí)際情況配置)
          limits:
            memory: 100Mi
          requests:
            memory: 100Mi

2, apply應(yīng)用YAML文件

[root@k8s-master1 ~]# kubectl apply -f daemonset-nginx.yml
daemonset.apps/daemonset-nginx created

3, 驗(yàn)證文章來(lái)源地址http://www.zghlxwxcb.cn/news/detail-503867.html

[root@master ~]# kubectl get daemonset				# daemonset可簡(jiǎn)寫為ds
[root@k8s-master1 ~]# kubectl get ds
NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
daemonset-nginx   4         4         4       4            4           <none>          114s

[root@k8s-master1 ~]# kubectl get pods -o wide
NAME                    READY   STATUS    RESTARTS   AGE   IP               NODE          NOMINATED NODE   READINESS GATES
daemonset-nginx-94z6d   1/1     Running   0          6s    10.244.194.104   k8s-worker1   <none>           <none>
daemonset-nginx-hs9mk   1/1     Running   0          6s    10.244.135.206   k8s-master3   <none>           <none>
daemonset-nginx-jrcf5   1/1     Running   0          6s    10.244.159.167   k8s-master1   <none>           <none>
daemonset-nginx-sslpl   1/1     Running   0          6s    10.244.224.22    k8s-master2   <none>           <none>

k8s集群中每個(gè)節(jié)點(diǎn)都會(huì)運(yùn)行一個(gè)pod

到了這里，關(guān)于kubernetes核心概念 controller的文章就介紹完了。如果您還想了解更多內(nèi)容，請(qǐng)?jiān)谟疑辖撬阉鱐OY模板網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章，希望大家以后多多支持TOY模板網(wǎng)！