Hi, dear!
Good day to you!
This topic describes how to restore a container in the exited state.
Symptom
On the CSP management node, run the docker ps -a | grep -v mesos command to query all CSP containers. Some containers are in the Exited state.
The container status is Abnormal on the Resources > Edge Sites > Containers page of the portal.
Cause
This issue may occur when the operating system of the management node restarts continuously.
In the scenario where the Docker daemon node restarts continuously, some containers are not started after the first power-off. As a result, HasBeenStartedBefore is not set and then the Docker daemon is powered off again. After the second power-off restart, Docker determines whether to restart the container based on HasBeenStartedBefore of the container on the disk. HasBeenStartedBefore is not set. As a result, the container cannot be started.
The health check scripts of these containers do not process the Exited state. Therefore, the containers are not started.
Impact on Services
In a single-node system, login to the CSP portal and service startup are affected.
In a cluster, two or three management nodes are deployed, which affects login to the CSP portal and service startup. If this problem occurs on only one management node, there is no impact on functions.
Identification Method
Log in to the CSP management node as user cspexpert and switch to user root.
Run the docker ps -a | grep -v mesos command to query all CSP containers. Some containers are in the Exited state.
Emergency Operations
Restart the container in the Exited state and check whether the container can be recovered.
Run the docker ps -a|grep -v mesos command to check whether the service is started properly (healthy state). If yes, no further action is required. If the process is not started properly, contact the document author.
Note: The docker restart command is used to process the CSP container. Do not use this command to operate the IVS container. The IVS container names start with mesos.
Best wishes!
