Restarting an instance is a common approach for troubleshooting the Cloud GPU Service. Restarting a GPU instance is equivalent to restarting the operating system of a local PC.
Overview
Restart preparation: The instance cannot provide service during restart. Make sure the Cloud GPU Service has suspended service requests before you restart it.
Restart procedure: It is recommended to use the restart operation provided by the cloud platform to restart the instance instead of running the restart command in the instance (such as the restart command in Windows and the Reboot command in Linux).
Restart duration: Generally speaking, the restart operation only takes a few minutes.
Instance physical characteristics: Restarting an instance does not alter its physical characteristics. The public IP address, private IP address, and any stored data on the instance remain unchanged.
Billing information: Restarting an instance does not initiate a new billing cycle.
Restarting an Instance In the Console
Restarting a single instance: Select the instance that needs to be restarted and click Restart at the top of the list. Alternatively, in the operation column on the right, select More > Instance Status > Restart.
Restarting multiple instances: Select all instances that need to be restarted. At the top of the list, click Restart to batch restart the instances. Instances that cannot be restarted will display the reason.