Protect Peak-Hour Performance by Scheduling Rebalancing Windows - PolarDB - Alibaba Cloud - PolarDB

When you perform scaling operations by adding or removing storage nodes, the system automatically triggers a data rebalancing task. This task consumes CPU and I/O resources on storage nodes and can degrade the performance of your business applications. To avoid performance degradation during peak hours, PolarDB-X lets you configure a maintenance window to schedule data rebalancing tasks for off-peak periods.

Prerequisites

Your instance version must be 5.4.18-17152226 or later.
This feature is available only for Enterprise Edition instances.

Note

For information about the instance version naming rules, see Release notes.
For information about how to view the version of an instance, see View and update the engine version of a PolarDB for Xscale instance.

Procedure

Log on to the PolarDB-X console.
In the top navigation bar, select the region where the instance is deployed.
On the Instances page, click the PolarDB-X 2.0 tab.
Find the instance that you want to manage and click its ID.

In the left-side navigation pane, click Configuration and management > Parameter settings. Set the compute layer parameters REBALANCE_MAINTENANCE_TIME_START and REBALANCE_MAINTENANCE_TIME_END.

Parameter descriptions:

Parameter	Description
REBALANCE_MAINTENANCE_TIME_START	Specifies the start time of the maintenance window for data rebalancing, in 24-hour format (HH:mm). Default value: `00:00`. Any pending or paused data rebalancing task starts or resumes at this time.
REBALANCE_MAINTENANCE_TIME_END	Specifies the end time of the maintenance window for data rebalancing, in 24-hour format (HH:mm). Default value: `00:00`. A data rebalancing task still running at this time is automatically paused and will resume during the next maintenance window.

Important

Parameter changes take effect immediately.
The system automatically splits a data rebalancing task into multiple subtasks based on the data volume. The maintenance window must be longer than the execution time of a subtask. Otherwise, a subtask might be repeatedly restarted but never complete, impeding the overall task progress. As a guideline, we recommend a maintenance window longer than 5 hours. If you have questions about setting the maintenance window duration, you can contact us.
After you set the parameters, if a data rebalancing task is running and the current time is outside the newly configured window, the task is immediately paused. It will resume during the next scheduled maintenance window.
If REBALANCE_MAINTENANCE_TIME_START is earlier than REBALANCE_MAINTENANCE_TIME_END, the maintenance window is from REBALANCE_MAINTENANCE_TIME_START to REBALANCE_MAINTENANCE_TIME_END on the same day.
If REBALANCE_MAINTENANCE_TIME_START is later than REBALANCE_MAINTENANCE_TIME_END, the maintenance window spans from REBALANCE_MAINTENANCE_TIME_START on the current day to REBALANCE_MAINTENANCE_TIME_END on the following day.
If REBALANCE_MAINTENANCE_TIME_START is the same as REBALANCE_MAINTENANCE_TIME_END, the maintenance window feature is disabled. In this case, after you add or remove a storage node, the data rebalancing task starts immediately, regardless of the business workload, and runs until completion.