【Kubernetes】创建Pod并分配到指定节点

一、编辑yaml文件###

[root@K8s-Master Tools]# cat hello-world-pod.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: hello-world
spec:                 
  nodeSelector:
    type: node1
  containers:
  - name: hello
    image: "ubuntu:14.04"
    command: [ "/bin/bash", "-ce", "tail -f /dev/null" ]

二、创建并查看运行状态###

[root@K8s-Master Tools]# kubectl create -f hello-world-pod.yaml 
pod/hello-world created
[root@K8s-Master Tools]# kubectl get pod hello-world
NAME          READY   STATUS    RESTARTS   AGE
hello-world   1/1     Running   0          8m2s
[root@K8s-Master Tools]# kubectl get pods --all-namespaces
NAMESPACE     NAME                                    READY   STATUS    RESTARTS   AGE
default       hello-world                             1/1     Running   0          3s
kube-system   coredns-bccdc95cf-plst5                 1/1     Running   13         36d
kube-system   coredns-bccdc95cf-vkqzr                 1/1     Running   13         36d
kube-system   etcd-k8s-master                         1/1     Running   15         36d
kube-system   kube-apiserver-k8s-master               1/1     Running   15         36d
kube-system   kube-controller-manager-k8s-master      1/1     Running   15         36d
kube-system   kube-flannel-ds-amd64-4g6qc             1/1     Running   6          36d
kube-system   kube-flannel-ds-amd64-ccj8v             1/1     Running   14         36d
kube-system   kube-flannel-ds-amd64-fhrfj             1/1     Running   8          36d
kube-system   kube-proxy-jp6mp                        1/1     Running   13         36d
kube-system   kube-proxy-rkbcx                        1/1     Running   6          36d
kube-system   kube-proxy-rmkqm                        1/1     Running   8          36d
kube-system   kube-scheduler-k8s-master               1/1     Running   15         36d
kube-system   kubernetes-dashboard-6db4897b74-lqfl6   1/1     Running   1          24h

注意点：将pod分配到指定的节点
将pod调度到指定节点有2种方式：
方式一：强制约束Pod调度到指定Node节点上

Pod.spec.nodeName

方式二：通过label-selector机制选择节点

Pod.spec.nodeSelector

以下使用第二种方式实现pod调度到指定节点。
①为Node添加label

kubectl label node k8s-node1 type=node1
kubectl get nodes --show-labels

[root@K8s-Master Tools]# kubectl get nodes --show-labels
NAME         STATUS   ROLES    AGE   VERSION   LABELS
k8s-master   Ready    master   37d   v1.15.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-master,kubernetes.io/os=linux,node-role.kubernetes.io/master=,type=master
k8s-node1    Ready    <none>   37d   v1.15.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node1,kubernetes.io/os=linux,type=node1
k8s-node2    Ready    <none>   37d   v1.15.3   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node2,kubernetes.io/os=linux,type=node2

②将pod分配到带有指定label的node

Pod操作：
删除Pod

kubectl delete pod hello

更新Pod

kubctl replace hello-world.yaml

三、可能遇到的问题###

Error：Back-off restarting failed container

[root@K8s-Master Tools]# kubectl describe pod hello-world
Name:         hello-world
Namespace:    default
Priority:     0
Node:         k8s-node1/10.0.2.15
Start Time:   Fri, 18 Oct 2019 17:16:13 +1100
Labels:       <none>
Annotations:  <none>
Status:       Running
IP:           192.168.1.15
Containers:
  hello:
    Container ID:  docker://ae85d78070dc02a19fed619c6270732cd98bfc1d824262b2935bc2362d7b70e5
    Image:         ubuntu:14.04
    Image ID:      docker-pullable://ubuntu@sha256:2f7c79927b346e436cc14c92bd4e5bd778c3bd7037f35bc639ac1589a7acfa90
    Port:          <none>
    Host Port:     <none>
    Command:
      /bin/echo
      Hello
      world
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Completed
      Exit Code:    0
      Started:      Fri, 18 Oct 2019 17:18:10 +1100
      Finished:     Fri, 18 Oct 2019 17:18:10 +1100
    Ready:          False
    Restart Count:  4
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-hctt8 (ro)
Conditions:
  Type              Status
  Initialized       True 
  Ready             False 
  ContainersReady   False 
  PodScheduled      True 
Volumes:
  default-token-hctt8:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-hctt8
    Optional:    false
QoS Class:       BestEffort
Node-Selectors:  type=node1
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:
  Type     Reason     Age                  From                Message
  ----     ------     ----                 ----                -------
  Normal   Scheduled  2m31s                default-scheduler   Successfully assigned default/hello-world to k8s-node1
  Normal   Pulling    2m30s                kubelet, k8s-node1  Pulling image "ubuntu:14.04"
  Normal   Pulled     2m10s                kubelet, k8s-node1  Successfully pulled image "ubuntu:14.04"
  Normal   Created    34s (x5 over 2m10s)  kubelet, k8s-node1  Created container hello
  Normal   Started    34s (x5 over 2m9s)   kubelet, k8s-node1  Started container hello
  Normal   Pulled     34s (x4 over 2m8s)   kubelet, k8s-node1  Container image "ubuntu:14.04" already present on machine
  Warning  BackOff    34s (x9 over 2m7s)   kubelet, k8s-node1  Back-off restarting failed container

解决方法：
As per Describe Pod command listing, your Container inside the Pod has been already completed with exit code 0, which states about successful completion without any errors/problems, but the life cycle for the Pod was very short. To keep Pod running continuously you must specify a task that will never finish.

apiVersion: v1
kind: Pod
metadata:
  name: ubuntu
spec:
  containers:
  - name: ubuntu
    image: ubuntu
    command: [ "/bin/bash", "-ce", "tail -f /dev/null" ]

参考链接：
将pod分配到指定的节点.
Back-off restarting failed container.