AWS ECS 容器无特定原因退出

Question

ECS 服务在运行几秒钟后退出。如果我们手动从镜像中创建一个容器，它运行良好，命令：sudo docker run -i -p 9100:9100 -p 9110:9110 -p 9120:9120 -p 9130:9130 847782638323.dkr.ecr.us -east-1.amazonaws.com/bytemark/cap

ECS日志如下：

2017-01-23T19:37:00Z [INFO] Created docker container for task bytemark-cap:2 arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e, Status: (CREATED->RUNNING) Containers: [bytemark-cap-container (CREATED->RUNNING),]: bytemark-cap-container(847782638323.dkr.ecr.us-east-1.amazonaws.com/bytemark/cap:latest) (CREATED->RUNNING) -> c8ee05c5cd688209939d96b54cb0c74c4122686036362d7bfa19f85bdc2dd56e
2017-01-23T19:37:00Z [INFO] Starting container module="TaskEngine" task="bytemark-cap:2 arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e, Status: (CREATED->RUNNING) Containers: [bytemark-cap-container (CREATED->RUNNING),]" container="bytemark-cap-container(847782638323.dkr.ecr.us-east-1.amazonaws.com/bytemark/cap:latest) (CREATED->RUNNING)"
2017-01-23T19:37:01Z [INFO] Task change event module="TaskEngine" event="{TaskArn:arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e Status:RUNNING Reason: SentStatus:NONE}"
2017-01-23T19:37:01Z [INFO] Adding event module="eventhandler" change="ContainerChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e bytemark-cap-container -> RUNNING, Ports [{9100 9100 0.0.0.0 0} {9110 9110 0.0.0.0 0} {9120 9120 0.0.0.0 0} {9130 9130 0.0.0.0 0}], Known Sent: NONE"
2017-01-23T19:37:01Z [INFO] Adding event module="eventhandler" change="TaskChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e -> RUNNING, Known Sent: NONE"
2017-01-23T19:37:01Z [INFO] Sending container change module="eventhandler" event="ContainerChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e bytemark-cap-container -> RUNNING, Ports [{9100 9100 0.0.0.0 0} {9110 9110 0.0.0.0 0} {9120 9120 0.0.0.0 0} {9130 9130 0.0.0.0 0}], Known Sent: NONE" change="ContainerChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e bytemark-cap-container -> RUNNING, Ports [{9100 9100 0.0.0.0 0} {9110 9110 0.0.0.0 0} {9120 9120 0.0.0.0 0} {9130 9130 0.0.0.0 0}], Known Sent: NONE"
2017-01-23T19:37:01Z [INFO] Redundant container state change for task bytemark-cap:2 arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e, Status: (RUNNING->RUNNING) Containers: [bytemark-cap-container (RUNNING->RUNNING),]: bytemark-cap-container(847782638323.dkr.ecr.us-east-1.amazonaws.com/bytemark/cap:latest) (RUNNING->RUNNING) to RUNNING, but already RUNNING
2017-01-23T19:37:01Z [INFO] Sending task change module="eventhandler" event="TaskChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e -> RUNNING, Known Sent: NONE" change="TaskChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e -> RUNNING, Known Sent: NONE"
2017-01-23T19:37:01Z [INFO] Task change event module="TaskEngine" event="{TaskArn:arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e Status:STOPPED Reason: SentStatus:RUNNING}"
2017-01-23T19:37:01Z [INFO] Error retrieving stats for container c8ee05c5cd688209939d96b54cb0c74c4122686036362d7bfa19f85bdc2dd56e: context canceled
2017-01-23T19:37:01Z [INFO] Adding event module="eventhandler" change="ContainerChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e bytemark-cap-container -> STOPPED, Exit 0, , Known Sent: RUNNING"
2017-01-23T19:37:01Z [INFO] Adding event module="eventhandler" change="TaskChange: arn:aws:ecs:us-east-1:847782638323:task/5d0c49a2-9591-469e-b05b-ebbaaefac26e -> STOPPED, Known Sent: RUNNING"
2017-01-23T19:37:01Z [INFO] Container c8ee05c5cd688209939d96b54cb0c74c4122686036362d7bfa19f85bdc2dd56e is terminal, stopping stats collection

有没有办法调试ECS服务退出的实际原因。

谢谢。

Answer 1

如果您可以通过 SSH 连接到 docker 实例（如果您通过 ECS 建立并指定了密钥或使用您自己的实例），您可以从 docker 获取日志。

我遇到了一个问题，我唯一的 docker 容器在几秒钟后就死机了，因此您可以使用

列出容器

docker ps -a

找到您要查看日志并使用的容器

docker logs *container id*