34 篇文章带有标签 “Kubernetes”

$ kubectl get pods
NAME          READY   STATUS              RESTARTS   AGE
kubia-j8tth   0/1     ContainerCreating   0          35s
kubia-jkt8g   1/1     Running             0          35s

$ kubectl describe rc kubia
Name:         kubia
Namespace:    kubia
Selector:     app=kubia
Labels:       app=kubia
Annotations:  Replicas:  2 current / 2 desired
Pods Status:  1 Running / 1 Waiting / 0 Succeeded / 0 Failed
Pod Template:
  Labels:  app=kubia
  Containers:
   kubia:
    Image:        wangjunjian/kubia:latest
    Port:         8080/TCP
    Host Port:    0/TCP
    Environment:  <none>
// ...

当删除 Pod 对象后，ReplicationController 会自动创建新的 Pod 对象来满足期望的副本个数。

$ kubectl delete pod kubia-jkt8g 
pod "kubia-jkt8g" deleted

• 编写 YAML 文件（kubia-namespace.yaml）

apiVersion: v1
kind: Namespace
metadata:
  name: kubia

$ kubectl apply -f kubia-namespace.yaml 
namespace/kubia created

查看新创建的名字空间，ns 是 namespace 的缩写。

$ kubectl get ns kubia
NAME    STATUS   AGE
kubia   Active   35s

• kubectl describe

$ kubectl describe pod kubia-manual

Name:         kubia-manual
Namespace:    default
Annotations:  gouchicao.com/someannotation: test

通过添加两个标签将 Pod 组织为两个维度（基于应用的横向维度和基于版本的纵向维度）

• 更新（--overwrite）

$ kubectl label pod kubia-manual-v2 app=kubia --overwrite
pod/kubia-manual-v2 labeled

$ kubectl get pods -L app,rel
NAME                    READY   STATUS    RESTARTS   AGE    APP     REL
kubia-864465c9d-744qc   1/1     Running   0          5d4h   kubia   
kubia-864465c9d-kzql7   1/1     Running   0          5d4h   kubia   
kubia-manual            1/1     Running   0          19h    kubia   
kubia-manual-v2         1/1     Running   0          22m    kubia   stable

删除（key-） $ kubectl label pod kubia-manual-v2 rel- pod/kubia-manual-v2 labeled $ kubectl get pods -L app,rel NAME READY STATUS RESTARTS AGE APP REL kubia-864465c9d-744qc 1/1 Running 0 5d4h kubia kubia-864465c9d-

通过端口转发（port-forward）可以连接到 Pod，方便测试和调试服务。

• JSON 格式

$ kubectl get pod kubia-manual -o json

$ kubectl logs kubia-manual 
Kubia server starting...

可以看到容器的ID作为主机名

如果在集群外访问，将localhost替换为集群中的任意一台服务器的IP即可。

• 方法2

openssl x509 -in /etc/kubernetes/pki/apiserver.crt -noout -text|grep ' Not '

            Not Before: Jun 16 06:09:07 2020 GMT
            Not After : Jun 16 06:09:07 2021 GMT

1. 添加策略配置文件

- --policy-config-file=/etc/kubernetes/scheduler-policy-config.json

2. 将卷挂载添加到Pod

- mountPath: /etc/kubernetes/scheduler-policy-config.json
  name: scheduler-policy-config
  readOnly: true

- hostPath:
      path: /etc/kubernetes/scheduler-policy-config.json
      type: FileOrCreate
  name: scheduler-policy-config

最终修改为 apiVersion: v1 kind: Pod metadata: creationTimestamp: null labels: component: kube-scheduler tier: control-plane name: kube-scheduler namespace: kube-system spec: containers: - command: - kube-scheduler - --authentication-kubeconfig=/etc/kubernetes/scheduler.

2. 重启服务

sudo systemctl restart docker

2. 使用Helm安装

helm install --generate-name nvdp/nvidia-device-plugin

失败（gpu2节点的Docker没有配置好） $ kubectl logs -n kube-system nvidia-device-plugin-1614240442-wfh6c 2021/02/26 07:03:48 Loading NVML 2021/02/26 07:03:48 Failed to initialize NVML: could not load NVML library. 2021/02/26 07:03:48 If this is a GPU node, did you set the docker default runtime to nvidia? 2021/02/26 07:03:48 You can check the prerequisites at: https://github.com/NVIDIA/k8s-device-plugin#prerequisites 2021/02/26 07:03:48 You can learn how to set the runtime at: https://github.