【Kubernetes】容器重启策略-OnFailure

环境

1. kubernetes 1.20.2
2. Spring Boot 2.5.0-M1

目标

创建一个 Pod 将其容器的重启策略设置为：OnFailure，查看容器停止时，Pod 的行为。

pod.yaml

直接创建 Pod 时，如果 restartPolicy 不写，默认为：Always。

apiVersion: v1
kind: Pod
metadata:
  name: pod-demo

spec:
  restartPolicy: OnFailure
  containers:
    - name: pod-demo
      image: jiangbo920827/spring-demo:actuator
      ports:
        - containerPort: 8080

查看 Pod

[root@master pod]# kubectl get -f pod.yaml -o wide
NAME       READY   STATUS    RESTARTS   AGE     IP            NODE    NOMINATED NODE   READINESS GATES
pod-demo   1/1     Running   0          2m24s   10.244.1.18   node1   <none>           <none>

可以看到 Pod 正常运行，RESTARTS（重启次数）字段为 0。

停止容器

非正常停止

[root@master pod]# kubectl exec pod-demo -- kill 1
[root@master pod]# kubectl get -f pod.yaml -o  wide
NAME       READY   STATUS    RESTARTS   AGE     IP            NODE    NOMINATED NODE   READINESS GATES
pod-demo   1/1     Running   1          3m51s   10.244.1.18   node1   <none>           <none>

可以看到容器被终止了，并且重启次数增加了一次，Pod 恢复到了正常状态。

正常停止

[root@master pod]# curl -X POST 10.244.1.18:8080/actuator/shutdown
{"message":"Shutting down, bye..."}[root@master pod]#
[root@master pod]# kubectl get -f pod.yaml -o  wide
NAME       READY   STATUS      RESTARTS   AGE     IP            NODE    NOMINATED NODE   READINESS GATES
pod-demo   0/1     Completed   1          4m50s   10.244.1.18   node1   <none>           <none>

状态变成了 Completed，READY 字段也变成了 0/1，也没有再次重启。

总结

介绍了容器的重启策略-OnFailure，在创建单个 Pod 的情况下，只有 Pod 在非正常情况下停止，才会重启。
容器的重启有一个回退策略，并以指数级增加。如果在 10 分钟没有错误，则重置重启计时。
详细说明见附录摘抄的官网说明。

附录

Container restart policy

The spec of a Pod has a restartPolicy field with possible values Always, OnFailure, and Never. The default value is Always.

The restartPolicy applies to all containers in the Pod. restartPolicy only refers to restarts of the containers by the kubelet on the same node. After containers in a Pod exit, the kubelet restarts them with an exponential back-off delay (10s, 20s, 40s, …), that is capped at five minutes. Once a container has executed for 10 minutes without any problems, the kubelet resets the restart backoff timer for that container.