逝水流LF

在minikube中操作k8s

• 标签: k8s

Minikube可以实现一种轻量级的Kubernetes集群，通过在本地计算机上创建虚拟机并部署只包含单个节点的简单集群。

在mac下的minikube中，学习了如果安装minikube以及简单的minikube操作。本节中学习如何发布应用／访问应用／滚动更新应用和回滚，扩容／缩容操作。

假设已经安装了minikube和kubectl，也有预备好的测试镜像hello-node

获取集群信息

1
kubectl cluster-info

创建Deployment

容器应用运行在pod上，通过kubectl run命令创建Deployment来管理Pod。

1
2
kubectl run hello-node --image=dillonliang/hello-node --port=8080
--> deployment.apps "hello-node" created

查看Deployment

1
2
3
4
5
kubectl get deployments

-->
NAME         DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
hello-node   1         1         1            1           3d

也可以通过dashboard查看整个集群的运行情况，
启动minikube dashboard

1
minikube dashboard

kubernetes dashboard

点击doshboard里的部署选项，可以看到刚才创建的hello-node

创建Service

默认情况下，我们不能直接访问kubernetes集群里的pods，kubernetes提供了一种访问集群内pods的方式：Service

1
2
3
## create service
kubectl expose deployment hello-node --type=LoadBalancer --port=8080
--> service "hello-node" exposed

查看刚刚创建的service

1
2
3
4
5
6
kubectl get services

-->
NAME         TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)          AGE
hello-node   LoadBalancer   10.107.249.99   <pending>     8080:30788/TCP   3d
kubernetes   ClusterIP      10.96.0.1       <none>        443/TCP          3d

可以看到有2个service，其中kubernetes是集群默认的，hello-node是刚刚创建的

在minikube中，使用如下命令启动service

1
2
minikube service hello-node
--> 🎉  Opening kubernetes service default/hello-node in default browser...

我本机测试的时候访问的是http://192.168.64.5:30788/，
因为minikube是单机集群，可以通过minikube ip命令获取集群地址，在上面的service里暴露了30788端口

滚动更新

假设我们的hello-node升级了，重新打包成镜像后的版本是0.0.2，如何更新k8s里的应用呢？

查看k8s里的hello-node

1
kubectl describe pods

这里只有一个pod，我们关注输出结果里的Containers->hello-node->Image字段，这里是dillonliang/hello-node

然后我们更新k8s里的应用的镜像,这里也是通过kubectl命令：

1
kubectl set image deployment/hello-node hello-node=dillonliang/hello-node:0.0.2

这时我们通过get pods获取k8s里的pod情况，毕竟在k8s里，pod是调度的最小单元，应用的更新／回滚都是依靠pod

1
2
3
4
5
kubectl get pods
-->
NAME                          READY     STATUS              RESTARTS   AGE
hello-node-58cb7dd798-mwbmj   0/1       ContainerCreating   0          3d
hello-node-6b8c5879f8-f75r4   1/1       Terminating         0          3d

可以看到旧的pod已经中止，新的pod正在创建中，稍等一会再次get pods，新的pod的状态变为了Running。查看pod的描述

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
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
kubectl describe pods

-->
Name:               hello-node-58cb7dd798-mwbmj
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               minikube/192.168.64.5
Start Time:         Fri, 26 Jul 2019 07:43:40 +0800
Labels:             pod-template-hash=58cb7dd798
                    run=hello-node
Annotations:        <none>
Status:             Running
IP:                 x.x.x.x
Controlled By:      ReplicaSet/hello-node-58cb7dd798
Containers:
  hello-node:
    Container ID:   docker://2c915697b9a33dd31b8d668d64761c55362c20faa3a55555b5b479aefeea6fed
    Image:          dillonliang/hello-node:0.0.2
    Image ID:       docker-pullable://dillonliang/hello-node@sha256:224545db28d48c54fc2b5375c85be61198ed7186bb79981036b5ba9642d946a5
    Port:           8080/TCP
    Host Port:      0/TCP
    State:          Running
      Started:      Fri, 26 Jul 2019 07:44:00 +0800
    Ready:          True
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-6ll8c (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             True
  ContainersReady   True
  PodScheduled      True
Volumes:
  default-token-6ll8c:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-6ll8c
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Normal  Scheduled  3d    default-scheduler  Successfully assigned default/hello-node-58cb7dd798-mwbmj to minikube
  Normal  Pulling    3d    kubelet, minikube  Pulling image "dillonliang/hello-node:0.0.2"
  Normal  Pulled     3d    kubelet, minikube  Successfully pulled image "dillonliang/hello-node:0.0.2"
  Normal  Created    3d    kubelet, minikube  Created container hello-node
  Normal  Started    3d    kubelet, minikube  Started container hello-node

从上面的输出可以看到，pod里的hello-node镜像已经是dillonliang/node-node:0.0.2，说明pod已经更新了，最下面的Events里也依次描述了pod的更新流程。

再次访问 http://192.168.64.5:30788/ ，可以看到新的输出结果hello-node-58cb7dd798-mwbmj: Hello World!，这里的hello-node-58cb7dd798-mwbmj是pod的hostname

回滚

实际情况中，我们的应用可能出现问题，需要回滚到之前的状态。

假设我们更新了image:0.0.10

1
kubectl set image deployment/hello-node hello-node=dillonliang/hello-node:0.0.10

然后我们通过kubectl get deployments查看到没有deployment是available，通过get pods查看到，新的镜像status是ErrImagePull，然后通过describe pods，发现Events里是找不到这个dillonliang/hello-node:0.0.10镜像

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
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
kubectl get deployments

-->
NAME         DESIRED   CURRENT   UP-TO-DATE   AVAILABLE   AGE
hello-node   1         1         1            0           3d


$ kubectl get pods

-->
NAME                          READY     STATUS         RESTARTS   AGE
hello-node-58cb7dd798-mwbmj   1/1       Terminating    0          3d
hello-node-69d57cc549-qpv5j   0/1       ErrImagePull   0          3d


$ kubectl describe pods

-->
Name:                      hello-node-58cb7dd798-mwbmj
Namespace:                 default
Priority:                  0
PriorityClassName:         <none>
Node:                      minikube/192.168.64.5
Start Time:                Fri, 26 Jul 2019 07:43:40 +0800
Labels:                    pod-template-hash=58cb7dd798
                           run=hello-node
Annotations:               <none>
Status:                    Terminating (lasts 3d)
Termination Grace Period:  30s
IP:                        x.x.x.x
Controlled By:             ReplicaSet/hello-node-58cb7dd798
Containers:
  hello-node:
    Container ID:   docker://2c915697b9a33dd31b8d668d64761c55362c20faa3a55555b5b479aefeea6fed
    Image:          dillonliang/hello-node:0.0.2
    Image ID:       docker-pullable://dillonliang/hello-node@sha256:224545db28d48c54fc2b5375c85be61198ed7186bb79981036b5ba9642d946a5
    Port:           8080/TCP
    Host Port:      0/TCP
    State:          Terminated
      Reason:       Error
      Exit Code:    137
      Started:      Fri, 26 Jul 2019 07:44:00 +0800
      Finished:     Fri, 26 Jul 2019 10:07:59 +0800
    Ready:          False
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-6ll8c (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  default-token-6ll8c:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-6ll8c
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type    Reason   Age   From               Message
  ----    ------   ----  ----               -------
  Normal  Killing  3d    kubelet, minikube  Stopping container hello-node


Name:               hello-node-69d57cc549-qpv5j
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               minikube/192.168.64.5
Start Time:         Fri, 26 Jul 2019 10:07:28 +0800
Labels:             pod-template-hash=69d57cc549
                    run=hello-node
Annotations:        <none>
Status:             Pending
IP:                 x.x.x.x
Controlled By:      ReplicaSet/hello-node-69d57cc549
Containers:
  hello-node:
    Container ID:
    Image:          dillonliang/hello-node:0.0.10
    Image ID:
    Port:           8080/TCP
    Host Port:      0/TCP
    State:          Waiting
      Reason:       ErrImagePull
    Ready:          False
    Restart Count:  0
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-6ll8c (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  default-token-6ll8c:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-6ll8c
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type     Reason     Age   From               Message
  ----     ------     ----  ----               -------
  Normal   Scheduled  3d    default-scheduler  Successfully assigned default/hello-node-69d57cc549-qpv5j to minikube
  Normal   Pulling    3d    kubelet, minikube  Pulling image "dillonliang/hello-node:0.0.10"
  Warning  Failed     3d    kubelet, minikube  Failed to pull image "dillonliang/hello-node:0.0.10": rpc error: code = Unknown desc = Error response from daemon: manifest for dillonliang/hello-node:0.0.10 not found
  Warning  Failed     3d    kubelet, minikube  Error: ErrImagePull
  Normal   BackOff    3d    kubelet, minikube  Back-off pulling image "dillonliang/hello-node:0.0.10"
  Warning  Failed     3d    kubelet, minikube  Error: ImagePullBackOff

这时候要进行回滚操作了，执行如下命令：

1
2
## 意思就是：取消deployment下hello-node的更新操作，即回滚。
kubectl rollout undo deployment/hello-node

再次通过get pods和describe pods命令查查hello-node的情况，发现回滚到之前的状态了。
值得注意的是，这次hostname变了，pod是一个新的pod。

动态扩容／缩容

线上有时候流量缓慢增高／减少，已有的pod会面临2中情况，流量增高时，pod压力变大，无法响应更多的请求，需要新增服务节点pod。流量减少的时候，pod会造成资源浪费，需要减少节点上的pod。

k8s提供了根据线上流量动态扩容／缩容的功能

1
kubectl scale deployments/hello-node --replicas=4

可以通过如下命令查看扩容状态：

1
2
3
kubectl get deployments
kubectl get pods -o wide
kubectl describe deployments/hello-node

此时，我们把pods扩容到4个，k8s为我们做了负载均衡，如何随机访问这4个pod呢？

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
kubectl describe services/hello-node

-->
Name:                     hello-node
Namespace:                default
Labels:                   run=hello-node
Annotations:              <none>
Selector:                 run=hello-node
Type:                     LoadBalancer
IP:                       10.107.249.99
Port:                     <unset>  8080/TCP
TargetPort:               8080/TCP
NodePort:                 <unset>  30788/TCP
Endpoints:                172.17.0.5:8080,172.17.0.6:8080,172.17.0.7:8080 + 1 more...
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>

使用如下命令来访问：

1
2
3
4
5
export NODE_PORT=$(kubectl get services/kubernetes-bootcamp -o go-template='{{(index .spec.ports 0).nodePort}}')

echo NODE_PORT=$NODE_PORT

curl $(minikube ip):$NODE_PORT

可以看到，每次随机访问4个pods中的一个。

如何缩容呢，修改–replicas=2就可以了，然后通过get deployments和get pods -o wide查看集群状态。

参考：
http://docs.kubernetes.org.cn/126.html
https://kubernetes.io/docs/tutorials/kubernetes-basics/