1. Power on and power off a node (hardware maintenance scenarios):

1. Before power-off: Make sure that no tasks run on the node and the node is not offline, and then enable Maintenance Mode for the node. The system will automatically migrate VMs and services on the node.

2. Power-off: After Maintenance Mode is enabled, power off the node. Do not directly unplug the power cord or press and hold the power button. After you complete hardware maintenance, power on the node, exit Maintenance Mode, and wait until data in the virtual storage is synced.
3. Verification: After data sync, migrate some VMs to the node to check whether services can run as expected.

2. Power on and power off a cluster (server room power cut or cluster migration scenarios):

1. Power off the cluster in the following procedure: Services > VMs > network devices > nodes > external networks. Shut down core services, such as databases and apps.

i. Shut down all VMs. Do not directly power off the VMs. Shut down the VMs on the console.

ii. Shut down virtual network devices, such as routers and switches.

iii. Shut down nodes other than the cluster controller and then shut down the cluster controller.

iv. Shut down external network devices, such as physical switches and storage devices.

2. Power on the cluster in the following procedure: External networks > nodes > network devices > VMs > services. Power on external network devices and test the connectivity.

i. Power on all nodes. Wait 10 to 20 minutes and then perform a health check. Make sure that the health check scores 100.

ii. Power on virtual network devices and verify the network status.

iii. Power on database VMs and then app VMs in sequence. Make sure that all services can run as expected.

3. Notes: Record the entire procedure in the operation log during power-on and power-off. If any anomaly occurs, such as a node failing to start, contact technical support immediately.