To rebalance the cluster load, you can rebalance the number of containers running on each node by moving the containers from nodes with high loads to the newly added nodes or nodes with low loads.
Log on to the Container Service console.
Click Services in the left-side navigation pane.
Select the cluster in which the service you want to reschedule resides from the Cluster list.
Click Reschedule at the right of the service you want to reschedule.
In the displayed dialog box, select or clear the Ignore Local Volumes and Force Reschedule check boxes. Then, click OK.
- Ignore Local Volumes: For containers with local volumes, rescheduling might move these containers to other machines and cause the data loss. To ignore the local volumes, select this check box. With this check box deselected, the containers with local volumes will not be rescheduled.
- Force Reschedule: Currently, to ensure the stability of online services, rescheduling only occurs when the memory usage and CPU usage of the machine are over 60% and 40% respectively by default. To ignore this limit, select this check box. Then, Container Service ignores the usage limit and forces rescheduling the service.
Note: The memory and CPU usage values are subject to the container configurations. Therefore, the values are not always the actual usage of the machine.