部署 pod 时出现 CrashBackloopError

Question

我是 kubernetes 的新手，正在尝试使用私有注册表部署一个 pod。每当我部署此 yaml 时，它都会崩溃循环。添加了 sleep 的大值想法，可能会导致此问题，但仍未奏效。

apiVersion: v1
kind: Pod
metadata:
  name: privetae-image-testing
spec:
  containers:
    - name: private-image-test
      image: buildforjenkin.azurecr.io/nginx:latest
      imagePullPolicy: IfNotPresent
      command: ['echo','success','sleep 1000000']

这里是日志：

Name:         privetae-image-testing
Namespace:    default
Priority:     0
Node:         docker-desktop/192.168.65.4
Start Time:   Sun, 24 Oct 2021 15:52:25 +0530
Labels:       <none>
Annotations:  <none>
Status:       Running
IP:           10.1.1.49
IPs:
  IP:  10.1.1.49
Containers:
  private-image-test:
    Container ID:  docker://46520936762f17b70d1ec92a121269e90aef2549390a14184e6c838e1e6bafec
    Image:         buildforjenkin.azurecr.io/nginx:latest
    Image ID:      docker-pullable://buildforjenkin.azurecr.io/nginx@sha256:7250923ba3543110040462388756ef099331822c6172a050b12c7a38361ea46f
    Port:          <none>
    Host Port:     <none>
    Command:
      echo
      success
      sleep 1000000
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Completed
      Exit Code:    0
      Started:      Sun, 24 Oct 2021 15:52:42 +0530
      Finished:     Sun, 24 Oct 2021 15:52:42 +0530
    Ready:          False
    Restart Count:  2
    Environment:    <none>
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-ld6zz (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  kube-api-access-ld6zz:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   BestEffort
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type     Reason     Age                From               Message
  ----     ------     ----               ----               -------
  Normal   Scheduled  34s                default-scheduler  Successfully assigned default/privetae-image-testing to docker-desktop
  Normal   Pulled     17s (x3 over 33s)  kubelet            Container image "buildforjenkin.azurecr.io/nginx:latest" already present on machine
  Normal   Created    17s (x3 over 33s)  kubelet            Created container private-image-test
  Normal   Started    17s (x3 over 33s)  kubelet            Started container private-image-test
  Warning  BackOff    2s (x5 over 31s)   kubelet            Back-off restarting failed container

我正在 Windows 上的 docker-desktop 上运行集群。 TIA

Answer 1

注意您使用的是标准 nginx 映像？尝试删除您的 pod 并重新申请：

apiVersion: v1
kind: Pod
metadata:
  name: private-image-testing
  labels:
    run: my-nginx
spec:
  restartPolicy: Always
  containers:
  - name: private-image-test
    image: buildforjenkin.azurecr.io/nginx:latest
    imagePullPolicy: IfNotPresent
    ports:
    - containerPort: 80
      name: http

如果您的 pod 运行，您应该可以使用 kubectl exec -it private-image-testing -- sh 远程进入，然后使用 wget -O- localhost 应该会打印一条欢迎消息。如果仍然失败，请将kubectl logs -f -l run=my-nginx 的输出粘贴到您的问题中。

Answer 2

查看我的previous answer 以逐步了解启动容器后发生的情况。

您正在启动一些 nginx:latest 容器，其中的进程将永远运行，以避免退出主进程。然后你添加覆盖（我会引用大卫的话：打印成功和睡眠 1000000，并打印这些词，然后退出）。

您不是让您的容器一直运行以提供服务，而是通过使用sleep 1000000 完成该过程来明确地射击您的腿。

当然，您的命令将被执行并且容器将退出。检查那个。它以状态 0 正确退出并执行了 2 次。未来还会更多。

  Reason:       CrashLoopBackOff
  Last State:   Terminated
  Reason:       Completed
  Exit Code:    0
  Started:      Sun, 24 Oct 2021 15:52:42 +0530
  Finished:     Sun, 24 Oct 2021 15:52:42 +0530

如果你真的需要command: ['echo','success','sleep 1000000']，你需要好好思考