Readiness probe failed:connection refused

我的K8S集群在啟動一個POD的時候說死起不來，然后就報下面的錯誤

Events:
  Type     Reason     Age                    From                                       Message
  ----     ------     ----                   ----                                       -------
  Normal   Scheduled  6m52s                  default-scheduler                          Successfully assigned alpha-vup/alpha-aggregatofrgql to ip-10-xxx-xxx-247.ad.abc.com
  Normal   Started    6m19s (x2 over 6m51s)  kubelet,  ip-10-xxx-xxx-247.ad.abc.com  Started container
  Warning  Unhealthy  5m58s (x6 over 6m48s)  kubelet,  ip-10-xxx-xxx-247.ad.abc.com  Readiness probe failed: Get http://10.xxx.xxx.107:80/: dial tcp 10.xxx.xxx.107:80: connect: connection refused
  Warning  Unhealthy  5m50s (x6 over 6m40s)  kubelet, ip-10-xxx-xxx-247.ad.abc.com  Liveness probe failed: Get http://10.xxx.xxx.107:80/: dial tcp 10.xxx.xxx.107:80: connect: connection refused

 

折騰了好久，解決方案就是注釋掉，下面的行，然后你就可以helm upgrade了

$ cat templates/deployment.yaml
              #          livenessProbe:
              #            httpGet:
              #              path: /
              #              port: http
              #          readinessProbe:
              #            httpGet:
              #              path: /
              #              port: http

總結：Liveness 和readines這兩個健康探針在使用時，你需要確保你的K8S集群中確實這兩個服務