活性探测失败：OCI 运行时执行失败：执行失败：EOF：未知答案

【问题标题】：Liveness probe failed: OCI runtime exec failed: exec failed: EOF: unknown活性探测失败：OCI 运行时执行失败：执行失败：EOF：未知
【发布时间】：2020-07-06 09:26:14
【问题描述】：

pod 启动一次失败。可能是什么原因？容器启动晚于 LivenessProbe 被激活？

活性探针配置为：

livenessProbe:
    exec:
      command:
      - /usr/bin/commandA
      - --is-alive
    initialDelaySeconds: 5
    periodSeconds: 2
    successThreshold: 1
    failureThreshold: 3

来自 pod 的错误日志

$ kubectl describe po t1004-rcpsdl-79c455547b-kdtvk -n cran1

Containers:
xxxxxx:
......
Port: <none>
Host Port: <none>
State: Running
Started: Fri, 03 Jul 2020 06:01:26 +0000
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Fri, 03 Jul 2020 06:00:45 +0000
Finished: Fri, 03 Jul 2020 06:01:25 +0000
Ready: True
Restart Count: 1

Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled <unknown> default-scheduler Successfully assigned cran1/t1004-rcpsdl-79c455547b-kdtvk to controller-0
Normal Pulled 105m (x2 over 106m) kubelet, controller-0 Container image "xxxxxxxxxxxxxxx" already present on machine
Normal Created 105m (x2 over 106m) kubelet, controller-0 Created container rcpsdl
Normal Started 105m (x2 over 106m) kubelet, controller-0 Started container rcpsdl
Warning Unhealthy 92m kubelet, controller-0 Liveness probe failed: OCI runtime exec failed: exec failed: EOF: unknown

【问题讨论】：

检查 pod 被调度的节点上的 kubelet 日志
您还有其他 pod 正在运行吗？错误来自运行时，您可以检查节点的日志（来自 kubernentes 和系统上）吗？您使用的是哪个容器运行时？

标签： kubernetes containers

【解决方案1】：

尝试增加这些参数的时间间隔

initialDelaySeconds: 300
periodSeconds: 10
successThreshold: 1
failureThreshold: 5

参考：Original Post

希望它对你有用。

【讨论】：