compute container

This article summarizes how to deal with problems related to computing containers, so that you can quickly locate the cause and complete the repair.

Problem 1: The computing container startup is always in the startup state, especially the Kubernetes cluster

Check node logs

cd ~/zhiqingyun-agent/works
cat ${containerId}.log

Question 2: When querying the container started by the Kubernetes cluster, an error message of 500 is reported or it keeps loading.

Enter the Kubernetes cluster and query node logs

kubectl get pods -n zhiqingyun-space
kubectl logs -f zhiqingyun-spark-container-sql-${containerId}-${containerId}-${podId}-driver -n zhiqingyun-space

一般由于域名无法访问导致的
计算容器，资源类型选择自定义，添加域名映射关系

{
  "qing.host1.name": "zhiqingyun-host1",
  "qing.host1.value": "127.0.0.1",
  "qing.host2.name": "zhiqingyun-host2",
  "qing.host2.value": "127.0.0.1",
  "qing.host3.name": "zhiqingyun-host3",
  "qing.host3.value": "127.0.0.1"
}