Updated on 2026-08-24 GMT+08:00

Handling Abnormal DWS Clusters

Canceling the Read-only Status: A cluster in read-only status does not allow write operations. You can remove this status on the management console. Triggering the read-only status may be due to high disk usage, so you need to clean up the cluster data. For details, see Solution to High Disk Usage and Cluster Read-Only.
Performing a Primary/Standby Switchback: When a primary/standby switchover occurs in a cluster, the cluster performance deteriorates. Take a single node with two DNs as an example. If a node is faulty and a primary/standby switchover occurs, the number of primary DNs on the node changes from 2 to 3. If the cluster load is high, the DNs will contend CPUs, I/O, and network resources, which may cause performance deterioration. In this case, the cluster can still be operated, but its performance is not as good as when the cluster is in a balanced state. You can perform a primary/standby switchback to change the cluster status to Available.

You can cancel the read-only status of a cluster:

The read-only status can be canceled for version 1.7.2 or later.
In 8.2.0 and later versions, you can free up disk space by using DROP/TRUNCATE TABLE in a read-only cluster.

You can cancel the read-only status only when a cluster is read-only.
When a cluster is in read-only status, stop the write tasks to prevent data loss caused by used up disk space.
After the read-only status is canceled, clear the data as soon as possible to prevent the cluster from entering the read-only status again after a period of time.

Log in to the DWS console.
In the navigation pane on the left, choose Cluster > Cluster List. All clusters are displayed by default.
In the row containing the cluster whose cluster status is Read-only, click Cancel Read-only.
In the dialog box that is displayed, click OK to confirm and cancel the read-only status for the cluster.

You can perform a primary/standby switchback for a cluster.

Only 8.1.1.202 and later versions support primary/standby cluster restoration.
Cluster restoration interrupts services for a short period of time. The interruption duration depends on the service volume. You are advised to perform this operation during off-peak hours.

Log in to the DWS console.
Choose Cluster > Cluster List and locate the cluster whose load is unbalanced.
In the Cluster Status column of the cluster, click Fix under Unbalanced.
In the dialog box that is displayed, confirm that the service is in off-peak hours, and click Yes. A message will be displayed in the upper right corner, indicating that the switchback request is being processed.
Check the cluster status. During the switchback, the cluster status is Switching back. After the switchback, the cluster status will change to Available.