Kubernetes Pod很快将状态从运行变为完成，我该如何防止这种情况发生

创建一个pod使用yaml，一旦创建了pod ，我正在kubectl exec运行我的gatling perf测试代码

kubectl exec gradlecommandfromcommandline -- ./gradlew gatlingRun-
simulations.RuntimeParameters -DUSERS=500 -DRAMP_DURATION=5 -DDURATION=30
但这是以kubectl控制台结束，并带有以下消息： -

命令以退出代码137终止

在调查中，它发现pod正在从运行阶段改变到完成阶段。

如何增加pod的生命周期，以便等待我的命令执行。这是pod yaml

apiVersion: v1
kind: Pod
metadata:
name: gradlecommandfromcommandline
labels:
purpose: gradlecommandfromcommandline
spec:
containers:

- name: gradlecommandfromcommandline
  image: tarunkumard/tarungatlingscript:v1.0
  workingDir: /opt/gatling-fundamentals/
  command: ["./gradlew"]
  args: ["gatlingRun-simulations.RuntimeParameters", "-DUSERS=500", "-

DRAMP_DURATION=5", "-DDURATION=30"]
restartPolicy: OnFailure

展开

收起

k8s小能手 2018-12-14 17:32:43 3232 0

2 条回答

写回答

取消提交回答

davidaks

在容器中，建议使用前台服务，而不是后台服务。比如运行 /bin/sshd, 而不是service sshd start.
这样可以让你的应用一直运行在前台，容器不会进入ready状态。
使用脚本也是一种方式，但是会有一个问题，容器的生命周期和应用不匹配，设计的时候可以考虑进去

2019-07-17 23:21:27

赞同 1 展开评论打赏
k8s小能手

整合最优质的专家资源和技术资料，问答解疑
这是yaml文件，使pod始终运行apiVersion：v1
kind: Pod
metadata:
name: gradlecommandfromcommandline
labels:
purpose: gradlecommandfromcommandline
spec:
volumes:
- name: docker-sock
  hostPath:
  path: /home/vagrant/k8s/pods/gatling/user-files/simulations # A file or
1. location on the node that you want to mount into the Pod
# command: [ "git clone https://github.com/TarunKDas2k18/PerfGatl.git" ]
containers:
- name: gradlecommandfromcommandline
  image: tarunkumard/tarungatlingscript:v1.0
workingDir: /opt/gatling-fundamentals/
command: ["./gradlew"]
args: ["gatlingRun-simulations.RuntimeParameters", "-DUSERS=500", "-
DRAMP_DURATION=5", "-DDURATION=30"]
- name: gatlingperftool
  image: tarunkumard/gatling:FirstScript # Run the ubuntu 16.04
command: [ "/bin/bash", "-c", "--" ] # You need to run some task inside a
container to keep it running
args: [ "while true; do sleep 10; done;" ] # Our simple program just sleeps inside
an infinite loop
volumeMounts:
```
- mountPath: /opt/gatling/user-files/simulations     # The mount path within the 
```
container
```
  name: docker-sock                   # Name must match the hostPath volume name
```
ports:
```
- containerPort: 80
```
2019-07-17 23:21:27

赞同 1 展开评论打赏