07 Deployment 功能及应用场景

发表于 | 分类于


Deployment 控制器

Pod与controllers的关系

  1. controllers: 在集群上管理和运行容器的对象
  2. 通过label-selector相关联
  3. Pod通过控制器实现应用的运维,如伸缩,滚动升级等
1
2
3
4
1. Pod是k8s当中最小的调度单元,但是一般很少直接使用pod
2. 使用控制器去对应用做部署
3. 控制器是高级对象，Pod是抽象的，控制器可以完成更高级的应用部署
4. controllers另外的学名 workload 工作负载

1
2
3
4
5
6
7
# 在一个yaml文件中可以分成两类:
1. 控制器
    - 指定控制器和api
    - 元数据
    - 标明标签，通过标签找到一组Pod 
2. Pod模板
    - 标明标签，关联控制器
1
2
3
# 知识点总结:
1. 控制器管理Pod,使用控制器去部署应用
2. 控制器通过标签关联Pod

Deployment 功能与应用场景

  1. 部署无状态应用
  2. 管理Pod和ReplicaSet
  3. 具有上线部署、副本设定、滚动升级、回滚等功能
  4. 提供声明式更新，例如值只更新一个新的Image
  5. 应用场景: Web服务，微服务
1
1. ReplicaSet也是控制器,但是他得功能比较单一

Deployment YAML文件字段解析

1
2
3
在k8s集群中创建pod的两种方式：
1. 使用命令创建 用于测试
2. 使用yaml文件 便于管理和复用

Deployment 部署无状态应用

1
2
3
[root@k8s-master1 demo]# kubectl create deployment --help
Usage:
  kubectl create deployment NAME --image=image [--dry-run] [options]
1
2
3
# 测试并输出到yaml格式文件
[root@k8s-master1 demo]# kubectl create deployment web --image=nginx1.16 --dry-run -o yaml
[root@k8s-master1 demo]# kubectl create deployment web --image=nginx1.16 --dry-run -o yaml > web.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
# 修改为应用属性
# 1. 副本个数
# 2. 标签名
# 3. 容器镜像

[root@k8s-master1 demo]# vim web.yaml 

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: web
  name: web
spec:
  replicas: 3
  selector:
    matchLabels:
      app: web
  template:
    metadata:
      labels:
        app: web
    spec:
      containers:
      - image: 172.17.70.252/project/java-demo:latest
        name: java-demo
1
2
3
4
5
6
7
8
9
10
11
[root@k8s-master1 demo]# kubectl get pods -o wide
NAME                   READY   STATUS    RESTARTS   AGE   IP            NODE        NOMINATED NODE   READINESS GATES
web-84c6fbbdcc-6tbkx   1/1     Running   0          28s   10.244.0.45   k8s-node1   <none>           <none>
web-84c6fbbdcc-ldv48   1/1     Running   0          28s   10.244.0.46   k8s-node1   <none>           <none>
web-84c6fbbdcc-qvl9f   1/1     Running   0          28s   10.244.1.48   k8s-node2   <none>           <none>

[root@k8s-master1 demo]# kubectl logs web-84c6fbbdcc-6tbkx 

[root@k8s-master1 demo]# kubectl get deploy
NAME   READY   UP-TO-DATE   AVAILABLE   AGE
web    3/3     3            3           6m23s

暴露应用

1
2
3
在k8s集群中暴露应用两种方式:
1. service
2. ingress
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
[root@k8s-master1 demo]# kubectl expose --help
Usage:
  kubectl expose (-f FILENAME | TYPE NAME) [--port=port] [--protocol=TCP|UDP|SCTP] [--target-port=number-or-name]
[--name=name] [--external-ip=external-ip-of-service] [--type=type] [options]

# kubectl expose service名称 pod资源 pod名称 service端口 pod端口 service类型
[root@k8s-master1 demo]# kubectl expose --name=web deployment web --port=80 --target-port=8080 --type=NodePort --dry-run -o yaml > web-service.yaml
[root@k8s-master1 demo]# vim web-service.yaml 

apiVersion: v1
kind: Service
metadata:
  labels:
    app: web
  name: web
spec:
  ports:
  - port: 80
    protocol: TCP
    targetPort: 8080
  selector:
    app: web                    # 关联pod标签
  type: NodePort

[root@k8s-master1 demo]# kubectl apply -f web-service.yaml 

[root@k8s-master1 demo]# kubectl get svc
NAME         TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)        AGE
kubernetes   ClusterIP   10.0.0.1     <none>        443/TCP        15d
web          NodePort    10.0.0.204   <none>        80:31619/TCP   13s

[root@k8s-master1 demo]# kubectl edit service/web

[root@k8s-master1 demo]# kubectl get ep
NAME         ENDPOINTS                                            AGE
kubernetes   172.17.70.251:6443,172.17.70.252:6443                16d
web          10.244.0.45:8080,10.244.0.46:8080,10.244.1.48:8080   3m50s
1
# nodeIP+端口测试访问

升级与回滚

升级操作 set

  1. 升级更新 
1
2
3
4
5
6
7
8
[root@k8s-master1 demo]# kubectl set --help
# 动态修改镜像
# 镜像通过tag版本号进行区分
[root@k8s-master1 demo]# kubectl set image --help
Usage:
  kubectl set image (-f FILENAME | TYPE NAME) CONTAINER_NAME_1=CONTAINER_IMAGE_1 ... CONTAINER_NAME_N=CONTAINER_IMAGE_N
[options]
# kubectl set image deployment/nginx busybox=busybox nginx=nginx:1.9.1
  1. 模拟程序升级
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
# 需要修改模拟程序并打包一个新的镜像,推送到镜像仓库,这里先用nginx替代

[root@k8s-master1 demo]# cat web.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: web
  name: web
spec:
  replicas: 3
  selector:
    matchLabels:
      app: web
  template:
    metadata:
      labels:
        app: web
    spec:
      containers:
      - image: 172.17.70.252/project/java-demo:latest
        name: java-demo

# deployment name
# 容器 name

[root@k8s-master1 demo]# kubectl set image deployment web java-demo=nginx:1.16
deployment.apps/web image updated

# 滚动升级 重新创建容器 失败了就不会删除之前的容器
# 先启动一个新镜像的pod 成功后再删除之前打好标记的pod 一个一个平滑升级
[root@k8s-master1 demo]# kubectl get pods
NAME                   READY   STATUS              RESTARTS   AGE
web-7bddf7b45f-6ntcs   0/1     ContainerCreating   0          4s
web-84c6fbbdcc-6tbkx   1/1     Running             0          30m
web-84c6fbbdcc-ldv48   1/1     Running             0          30m
web-84c6fbbdcc-qvl9f   1/1     Running             0          30m
1
2
3
4
5
6
7
8
9
10
# 错误原因 镜像名字写错了 
Events:
  Type     Reason     Age                 From                Message
  ----     ------     ----                ----                -------
  Normal   Scheduled  <unknown>           default-scheduler   Successfully assigned default/web-7bddf7b45f-6ntcs to k8s-node1
  Normal   BackOff    23s (x4 over 102s)  kubelet, k8s-node1  Back-off pulling image "nginx1.16"
  Warning  Failed     23s (x4 over 102s)  kubelet, k8s-node1  Error: ImagePullBackOff
  Normal   Pulling    11s (x4 over 110s)  kubelet, k8s-node1  Pulling image "nginx1.16"
  Warning  Failed     6s (x4 over 103s)   kubelet, k8s-node1  Failed to pull image "nginx1.16": rpc error: code = Unknown desc = Error response from daemon: pull access denied for nginx1.16, repository does not exist or may require 'docker login'
  Warning  Failed     6s (x4 over 103s)   kubelet, k8s-node1  Error: ErrImagePull
1
2
3
4
5
6
7
8
9
10
11
12
13
14
[root@k8s-master1 demo]# kubectl set image deployment web java-demo=nginx:1.16

# 还需要修改service里面的端口 8080 -> 80
[root@k8s-master1 demo]# kubectl edit service/web
...
spec:
  clusterIP: 10.0.0.204
  externalTrafficPolicy: Cluster
  ports:
  - nodePort: 31619
    port: 80
    protocol: TCP
    targetPort: 80                 # 修改目标端口
...
  1. 查看升级状态
1
2
# 如果pod做了健康检查,那么在更新的时候可以查看更新状态
[root@k8s-master1 demo]# kubectl rollout status deployment/web
  1. kubectl patch 补丁命令
1
[root@k8s-master1 demo]# kubectl patch --help

回滚操作 rollout

  1. 查看deployment历史记录
1
2
3
4
5
6
7
8
9
[root@k8s-master1 demo]# kubectl rollout --help

[root@k8s-master1 demo]# kubectl rollout history deployment/web
deployment.apps/web 
REVISION  CHANGE-CAUSE
1         <none>       # 第一次部署demo 
2         <none>
3         <none>
4         <none>
  1. 回滚到上一个版本
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
[root@k8s-master1 demo]# kubectl rollout undo deployment/web
deployment.apps/web rolled back

# 滚动回滚 部署了新的pod

[root@k8s-master1 demo]# kubectl get pods
NAME                   READY   STATUS        RESTARTS   AGE
web-68d86dd855-pgjlq   0/1     Terminating   0          13m
web-68d86dd855-pp8mv   0/1     Terminating   0          13m
web-68d86dd855-pq6w6   0/1     Terminating   0          13m
web-8598778cf-2m7lb    1/1     Running       0          9s
web-8598778cf-fnjtd    1/1     Running       0          8s
web-8598778cf-qjsd6    1/1     Running       0          10s
[root@k8s-master1 demo]# kubectl get pods
NAME                  READY   STATUS    RESTARTS   AGE
web-8598778cf-2m7lb   1/1     Running   0          14s
web-8598778cf-fnjtd   1/1     Running   0          13s
web-8598778cf-qjsd6   1/1     Running   0          15s
  1. 回滚到指定版本
1
2
3
4
5
6
# 我这个nginx 回滚到 java-demo  别忘记修改回端口

[root@k8s-master1 demo]# kubectl rollout undo deployment/web --to-revision=1
deployment.apps/web rolled back
[root@k8s-master1 demo]# kubectl edit svc/web
service/web edited

删除应用

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# 删除pod还会拉起,因为创建的是deployment会保证副本数量
[root@k8s-master1 demo]# kubectl delete pod web-84c6fbbdcc-7dh5h
pod "web-84c6fbbdcc-7dh5h" deleted
[root@k8s-master1 demo]# kubectl get pods
NAME                   READY   STATUS    RESTARTS   AGE
web-84c6fbbdcc-w2v5g   1/1     Running   0          3m33s
web-84c6fbbdcc-w56vs   1/1     Running   0          4s
web-84c6fbbdcc-xzglk   1/1     Running   0          3m30s

# 需要删除deployment
[root@k8s-master1 demo]# kubectl delete deployment/web
deployment.apps "web" deleted
[root@k8s-master1 demo]# kubectl get pods
NAME                   READY   STATUS        RESTARTS   AGE
web-84c6fbbdcc-w2v5g   0/1     Terminating   0          4m5s
web-84c6fbbdcc-w56vs   0/1     Terminating   0          36s
1
2
3
4
5
6
7
8
9
10
[root@k8s-master1 demo]# kubectl apply -f web.yaml 
deployment.apps/web created
[root@k8s-master1 demo]# kubectl rollout history deployment/web
deployment.apps/web 
REVISION  CHANGE-CAUSE
1         <none>

# service 和 pod 是不同的一种组件,删除pod并不会影响已创建的svc,这种解耦太棒了
# 删除svc
[root@k8s-master1 demo]# kubectl delete svc/web

弹性伸缩 scale

1
2
1. 根据活动扩容实例
2. 传统的方式是虚拟机,准备扩容环境
1
2
3
4
5
6
7
8
9
10
11
12
13
# kubectl scale 快速扩容 承担并发
[root@k8s-master1 demo]# kubectl scale --help

[root@k8s-master1 demo]# kubectl scale --replicas=5 deployment/web 
deployment.apps/web scaled

[root@k8s-master1 demo]# kubectl get pods -o wide
NAME                   READY   STATUS    RESTARTS   AGE     IP            NODE        NOMINATED NODE   READINESS GATES
web-84c6fbbdcc-685nc   1/1     Running   0          5m41s   10.244.0.55   k8s-node1   <none>           <none>
web-84c6fbbdcc-88289   1/1     Running   0          30s     10.244.1.57   k8s-node2   <none>           <none>
web-84c6fbbdcc-9mdjv   1/1     Running   0          5m41s   10.244.0.56   k8s-node1   <none>           <none>
web-84c6fbbdcc-dvdkn   1/1     Running   0          5m41s   10.244.1.55   k8s-node2   <none>           <none>
web-84c6fbbdcc-fpwqt   1/1     Running   0          30s     10.244.1.56   k8s-node2   <none>           <none>
  1. 如果集群的资源不够,单只扩容pod是不行的
1
2
3
4
1. 增加nodes节点或者增加nodes的资源
2. 如果扩容node节点 都是根据业务的量去扩容,并且提前规划
3. 现在都是手动扩容,缩容,如何自动弹性实现
4. 自动需要借助node的资源指标进行弹性扩容,比如cpu,内存等
  1. 手动缩容
1
2
3
4
5
6
7
8
9
10
11
12
13
14
[root@k8s-master1 demo]# kubectl scale --replicas=3 deployment/web 
deployment.apps/web scaled
[root@k8s-master1 demo]# kubectl get pods -o wide
NAME                   READY   STATUS        RESTARTS   AGE     IP            NODE        NOMINATED NODE   READINESS GATES
web-84c6fbbdcc-685nc   1/1     Running       0          9m36s   10.244.0.55   k8s-node1   <none>           <none>
web-84c6fbbdcc-88289   0/1     Terminating   0          4m25s   10.244.1.57   k8s-node2   <none>           <none>
web-84c6fbbdcc-9mdjv   1/1     Running       0          9m36s   10.244.0.56   k8s-node1   <none>           <none>
web-84c6fbbdcc-dvdkn   1/1     Running       0          9m36s   10.244.1.55   k8s-node2   <none>           <none>
web-84c6fbbdcc-fpwqt   0/1     Terminating   0          4m25s   10.244.1.56   k8s-node2   <none>           <none>
[root@k8s-master1 demo]# kubectl get pods -o wide
NAME                   READY   STATUS    RESTARTS   AGE     IP            NODE        NOMINATED NODE   READINESS GATES
web-84c6fbbdcc-685nc   1/1     Running   0          9m41s   10.244.0.55   k8s-node1   <none>           <none>
web-84c6fbbdcc-9mdjv   1/1     Running   0          9m41s   10.244.0.56   k8s-node1   <none>           <none>
web-84c6fbbdcc-dvdkn   1/1     Running   0          9m41s   10.244.1.55   k8s-node2   <none>           <none>

Deployment 和 ReplicaSet

1
2
3
4
5
6
7
8
1. 部署1个Deployment,会自动创建1个ReplicaSet和Pod

[root@k8s-master1 demo]# kubectl get deploy,rs
NAME                  READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/web   3/3     3            3           12m

NAME                             DESIRED   CURRENT   READY   AGE
replicaset.apps/web-84c6fbbdcc   3         3         3       12m
1
2
3
4
5
1. 部署Deployment -> 创建RS -> RS管理多个Pod
2. RS会判断replicas: 3 副本数 也会获取当前Pod的副本数
3. 如果少了一个 RS会帮助我们拉起
4. RS 帮我们实现了控制副本个数 和 历史版本记录,rollout回滚找rs回滚
5. 滚动更新会创建新的RS
1
2
3
1. Deployment -> 创建RS -> RS管理多个Pod
2. 滚动更新 创建新的RS -> 启动新的Pod 关闭之前rs中的1个,直到全部更新完成
3. service 指向 pod不变
1
2
3
4
5
6
7
8
9
10
11
12
13
14
[root@k8s-master1 ~]# kubectl set image deployment/web java-demo=nginx:1.16

[root@k8s-master1 ~]# kubectl get rs,deploy
NAME                             DESIRED   CURRENT   READY   AGE
replicaset.apps/web-68d86dd855   3         3         3       17s
replicaset.apps/web-7bddf7b45f   0         0         0       2m
replicaset.apps/web-84c6fbbdcc   0         0         0       107m

[root@k8s-master1 ~]# kubectl rollout history deployment/web
deployment.apps/web 
REVISION  CHANGE-CAUSE
1         <none>
2         <none>
3         <none>