Dynamic Power Management (DPM) enables the system to automatically shut down or wake up hosts based on the monitoring policies. When the load of the cluster decreases, the system automatically shuts down the hosts with low resource usage after migrating all VMs on the hosts to other hosts in the cluster. When the load of the cluster increases, the system wakes up the shutdown hosts and distributes load to the hosts through DRS.
Operators can configure host reclaim conditions and host wakeup conditions. When the conditions are met, the system automatically shuts down or wakes up hosts. The host reclaim conditions or host wakeup conditions include CPU usage and memory usage criteria. Operators can configure the system to perform DPM when the CPU usage or memory usage exceeds the set threshold or both of the conditions exceed their thresholds. For more information about monitoring policies, see "Manage monitoring policites."
Host reclaim—When the CPU usage and memory usage of a host remain lower than the threshold specified in the reclaim policy for the specified time period, the system automatically shuts down the host after migrating all VMs on the host to other hosts in the cluster. The system stops shutting down hosts when the cluster has only two or less hosts that are operating correctly.
Host wakeup—When the CPU usage and memory usage of normal hosts remain higher than the thresholds specified in the wakeup policy for the specified time period, the system automatically wakes up hosts in shutdown state and migrates some VMs on the overloaded host to the hosts that have been woken up.
To enable intelligent power management, make sure all hosts and VMs in the cluster use the same shared storage and the VMs are enabled with automatic migration.
As a best practice, leave an appropriate gap between the thresholds specified in the wakeup and reclaim policies. For example, you can specify the condition CPU/memory usage < 10% in the reclaim policy, and specify the condition CPU/memory usage > 80% in the wakeup policy.
The monitoring policies for DPM cannot use the connection number as a condition.
On the top navigation bar, click Resources.
From the left navigation pane, select Compute > Host Pool Name > Cluster Name.
Click More, and then select DPM.
Enable DPM, specify the duration and detection interval, and select a reclaim policy and a wakeup policy.
Click OK.
Enable DPM: Select whether to enable DPM.
Duration: Specify the time to wait for DPM to take effect after the CPU or memory usage reaches or exceeds the threshold specified in the reclaim or wakeup policy.
Detection Interval: Specify the interval at which the system checks the CPU usage and memory usage of the hosts.
Reclaim Policy: Select a host reclaim policy. When the load of a host remains lower than the threshold specified in the reclaim policy for the specified time period, the system automatically shuts down the host after migrating all VMs on the host to the host with the lowest memory usage.
Wakeup Policy: Select a host wakeup policy. When the load of a host remains higher than the threshold specified in the wakeup policy for the specified time period, the system automatically wakes up hosts in shutdown state and migrates some VMs on the overloaded host to the hosts that have been woken up.
Time Range: Specify a time range during which dynamic power management takes effect.
If the start time is earlier than the end time, the time range takes effect on the day of configuration.
If the start time is the same as the end time, dynamic power management takes effect all the day.