Scenario

When the server room is undergoing routine maintenance (routine power outage) or relocation, it is necessary to power off and power on the HCI cluster. To fully guarantee the service system's availability and avoid the power-off process's impact on the user's service data. You should manually shut down business services before powering off, shut down the virtual machine, and finally shut down the HCI cluster.

Precautions

1. Before powering off the HCI cluster, ensure that no other tasks are being executed in the entire cluster.
2. The virtual machine's shutdown operation should use the HCI web console's shutdown function or the virtual machine's shutdown function. It is forbidden to use the power-off function of the HCI web console.
3. The shutdown operation of the server should be performed on the node page of the HCI web console. It is forbidden to directly unplug the power supply or press and hold the server shutdown button to perform the shutdown operation.
4. The cluster power-off process should follow the sequence of shutting down services first, then shutting down virtual machines, then servers, and finally network devices.
5. The cluster power-on process should follow the sequence of turning on network devices, turning on servers, turning on virtual machines, and finally turning on business services.

Steps

Cluster power-off process

1. Log in to the HCI web console and check whether any tasks are being executed in the task list of the HCI platform. If a task is being executed, wait for the task to finish running before continuing the operation.

2. Manually shut down the service of the business system (optional)
3. On the Compute > Sangfor HCI page, select all virtual machines, select the shutdown operation, and ensure that all virtual machines are powered off.

4. Navigate the Networking > Topology, and click Devices to enter the device list page.

5. Select all devices and click Shut Down to shut down the virtual network device.

6. After the shutdown is completed, you can use the method in step 1 to check whether tasks are being executed in the task list.
7. Open the Nodes page, select the non-cluster controller nodes one by one, and click More > Power Off to perform the shutdown operation. Finally, select the cluster controller and click More > Power Off to perform the shutdown operation.

8. After confirming that the server is shut down, shut down the associated network devices.

Cluster power-on process

1. Power on the network device and perform a network test to ensure the network is normal.
2. Power up the server and ensure that all power to the server is functioning properly.
3. After the server starts normally, wait for 10-20 minutes, log in to the HCI console, and perform Health Check on the home page to ensure that the Health Check score is 100 points.

4. Check whether tasks are being executed in the task list of the HCI platform. If there are tasks to be executed, you need to wait for the tasks to be executed before continuing.

5. Navigate the Networking > Topology, and click Devices to enter the device list page.

6. Select all devices and click Power On to start the virtual network device.

7. On the Virtual Machine page, start the virtual machine step by step, and enable the services of the production system on the virtual machine. The power on and enabling services should follow the sequence of powering on the database virtual machine first and then powering on the application virtual machine after the database service is normal.

8. During the startup process, verify whether the business system is normal one by one. After the current business system is normal, start other business systems.