分布式提交报错

container日志报错信息如下：

2019-04-24 18:17:01,747 INFO xdl.ContainerBase: pull image [docker pull registry.cn-hangzhou.aliyuncs.com/xdl/xdl:ubuntu-cpu-mxnet1.3]
2019-04-24 18:17:02,533 INFO xdl.ContainerBase: cmd [docker pull registry.cn-hangzhou.aliyuncs.com/xdl/xdl:ubuntu-cpu-mxnet1.3] cost 786 ms status [0]
2019-04-24 18:17:02,534 INFO xdl.ContainerBase: pull image cost 787 ms
2019-04-24 18:17:04,165 INFO xdl.ContainerRunner$ContainerSignalHandler: Container is killed by signal:15.
2019-04-24 18:17:04,165 INFO xdl.ContainerBase: stop container cmd [docker stop -t 30 xdl_application_1556099741086_0003_scheduler_0_000002]

以上没有明确报错信息，只提示killed，请教下是什么原因呢？

Start by reading the container lifecycle around xdl.ContainerBase and xdl.ContainerRunner$ContainerSignalHandler, then inspect the distributed submission and container logs surrounding the SIGTERM. Reproduce the submission if possible and determine what initiates the docker stop command; done means the cause of the termination is identified and a verifiable fix or diagnostic is documented.

Written by the indexing model from the issue text.