AKS Pod CrashLoopBackOff 上的 Helm kubernetes

Question

我正在嘗試通過 Azure Kubernetes 服務上的 helm kubernetes 部署一個簡單的 nodejs 應用程序，但在拉取我的圖像后，它顯示CrashLoopBackOff 。

這是我迄今為止嘗試過的：

我的Dockerfile ：

FROM node:6
WORKDIR /usr/src/app
COPY package*.json ./
RUN npm install
COPY . .
EXPOSE 32000
CMD [ "npm", "start" ]

我的server.js ：

'use strict';

const express = require('express');

const PORT = 32000;
const HOST = '0.0.0.0';

const app = express();
app.get('/', (req, res) => {
  res.send('Hello world from container.\n');
});

app.listen(PORT, HOST);
console.log(`Running on http://${HOST}:${PORT}`);

我已將此圖像推送到 ACR。

新更新：這是kubectl describe pod POD_NAME的完整輸出：

Name:               myrel02-mychart06-5dc9d4b86c-kqg4n
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               aks-nodepool1-19665249-0/10.240.0.6
Start Time:         Tue, 05 Feb 2019 11:31:27 +0500
Labels:             app.kubernetes.io/instance=myrel02
                    app.kubernetes.io/name=mychart06
                    pod-template-hash=5dc9d4b86c
Annotations:        <none>
Status:             Running
IP:                 10.244.2.5
Controlled By:      ReplicaSet/myrel02-mychart06-5dc9d4b86c
Containers:
  mychart06:
    Container ID:   docker://c239a2b9c38974098bbb1646a272504edd2d199afa50f61d02a0ce335fe60660
    Image:          registry-1.docker.io/arycloud/docker-web-app:0.5
    Image ID:       docker-pullable://registry-1.docker.io/arycloud/docker-web-app@sha256:4faab280d161b727e0a6a6d9dfb52b22cf9c6cd7dd07916d6fe164d9af5737a7
    Port:           80/TCP
    Host Port:      0/TCP
    State:          Waiting
      Reason:       CrashLoopBackOff
    Last State:     Terminated
      Reason:       Completed
      Exit Code:    0
      Started:      Tue, 05 Feb 2019 11:39:56 +0500
      Finished:     Tue, 05 Feb 2019 11:40:22 +0500
    Ready:          False
    Restart Count:  7
    Liveness:       http-get http://:http/ delay=0s timeout=1s period=10s #success=1 #failure=3
    Readiness:      http-get http://:http/ delay=0s timeout=1s period=10s #success=1 #failure=3
    Environment:
      KUBERNETES_PORT_443_TCP_ADDR:  cluster06-ary-2a187a-dc393b82.hcp.centralus.azmk8s.io
      KUBERNETES_PORT:               tcp://cluster06-ary-2a187a-dc393b82.hcp.centralus.azmk8s.io:443
      KUBERNETES_PORT_443_TCP:       tcp://cluster06-ary-2a187a-dc393b82.hcp.centralus.azmk8s.io:443
      KUBERNETES_SERVICE_HOST:       cluster06-ary-2a187a-dc393b82.hcp.centralus.azmk8s.io
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from default-token-gm49w (ro)
Conditions:
  Type              Status
  Initialized       True
  Ready             False
  ContainersReady   False
  PodScheduled      True
Volumes:
  default-token-gm49w:
    Type:        Secret (a volume populated by a Secret)
    SecretName:  default-token-gm49w
    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  10m                    default-scheduler                  Successfully assigned default/myrel02-mychart06-5dc9d4b86c-kqg4n to aks-nodepool1-19665249-0
  Normal   Pulling    10m                    kubelet, aks-nodepool1-19665249-0  pulling image "registry-1.docker.io/arycloud/docker-web-app:0.5"
  Normal   Pulled     10m                    kubelet, aks-nodepool1-19665249-0  Successfully pulled image "registry-1.docker.io/arycloud/docker-web-app:0.5"
  Warning  Unhealthy  9m30s (x6 over 10m)    kubelet, aks-nodepool1-19665249-0  Liveness probe failed: Get http://10.244.2.5:80/: dial tcp 10.244.2.5:80: connect: connection refused
  Normal   Created    9m29s (x3 over 10m)    kubelet, aks-nodepool1-19665249-0  Created container
  Normal   Started    9m29s (x3 over 10m)    kubelet, aks-nodepool1-19665249-0  Started container
  Normal   Killing    9m29s (x2 over 9m59s)  kubelet, aks-nodepool1-19665249-0  Killing container with id docker://mychart06:Container failed liveness probe.. Container will be killed and recreated.
  Warning  Unhealthy  9m23s (x7 over 10m)    kubelet, aks-nodepool1-19665249-0  Readiness probe failed: Get http://10.244.2.5:80/: dial tcp 10.244.2.5:80: connect: connection refused
  Normal   Pulled     5m29s (x6 over 9m59s)  kubelet, aks-nodepool1-19665249-0  Container image "registry-1.docker.io/arycloud/docker-web-app:0.5" already present on machine
  Warning  BackOff    22s (x33 over 7m59s)   kubelet, aks-nodepool1-19665249-0  Back-off restarting failed container

更新： docker logs CONTAINER_ID輸出：

> nodejs@1.0.0 start /usr/src/app
> node server.js

Running on http://0.0.0.0:32000

我怎樣才能避免這個問題？

提前致謝！

Answer 1

正如我從kubectl describe pod命令輸出中看到的，Pod 內的 Container 已經完成，退出代碼為 0（@4c74356b41 在評論中提到了這一點）。 Reason: Completed ，說明成功完成，沒有任何錯誤/問題。 但是 Pod 的生命周期很短，因此 Kubernetes 會不斷調度新的 Pod，但是 Liveness 和 Readiness 探針仍然無法對容器的健康狀況進行檢測。

為了保持 Pod 運行，您必須在容器內部指定一個任務（進程），以便能夠持續運行。 關於如何解決此類問題，有很多討論和解決方案，可以在此處找到更多提示。

Answer 2

kubectl logs 命令僅在 pod 啟動並運行時才有效。 如果不是，您可以使用 kubectl events 命令。 它會給你一些事件日志，有時（根據我的經驗）還會給你關於正在發生的事情的線索。

kubectl get events -n <your_app_namespace> --sort-by='.metadata.creationTimestamp'

默認情況下，它不對事件進行排序，因此使用 --sort-by 標志。