HoloWeb provides a visual interface to change the table group of a table, a process known as resharding. This allows you to quickly perform resharding operations and handle any exceptions that may occur. This topic describes how to use HoloWeb to reshard tables. For more information about table groups, shard counts, and the principles of resharding, see Key concepts.
Limitations
-
HoloWeb supports resharding only for instances of V2.0.24 and later. If your instance is an earlier version, upgrade your instance or see Move a table to a new table group (resharding).
-
For other limitations of the resharding feature, see Move a table to a new table group (resharding).
Create a table resharding task
-
Go to the HoloWeb development page. For more information, see Connect to HoloWeb.
-
In the top navigation bar of the HoloWeb development page, click Data Solutions.
-
On the Data Solutions page, click Table Resharding in the navigation pane on the left.
-
On the Table Resharding page, click Create Table Resharding Task and configure the resharding task using the parameters in the following table. You can keep the default values for parameters that are not mentioned.
Section
Parameter
Description
Select Instance
Instance Name
The name of the target Hologres instance.
Database
The database in the target instance.
Table Information
Schema
The schema that contains the table to be resharded.
Table
The table to reshard.
Target Table Group
The destination table group.
Import GUC Parameter Settings
GUC Parameters
Optional. Resharding involves data import, which can cause high memory overhead. You may need to adjust GUC parameters to manage memory usage. For more information, see How do I resolve an out-of-memory (OOM) error during data import or export?.
-
Click Submit. You can then view the resharding tasks on the Table Resharding page. After the task is created, the Table Resharding task list displays the instance name, database, table name, original table group, target table group, status (such as Successful), and progress for each task. In the Operation column, you can click Details to view task details, click Delete to remove the task record, or click Execution History to view the execution history.
NoteProgress: For a partitioned table, the progress is calculated based on the number of completed child partitions. For a non-partitioned table, the progress is displayed only as 0% or 100%.
-
The resharding task is complete when its Status is Successful.
Troubleshooting
During the resharding process, the system first sets the table to a read-only state and then moves it to the target table group by using a temporary table. Because resharding involves data import, the task can be time-consuming and prone to exceptions, such as an OOM error or manual termination. In these cases, the task's Status is set to Interrupted or Failed. When a resharding task is interrupted or fails, the original table remains in a read-only state, and a temporary table named <initial_table_name>_xxxxxxxx remains in the database. You must then choose one of the following actions:
-
Resume task
If the task was accidentally interrupted, or if it failed due to an issue like an OOM error, you can resume it after adjusting the GUC parameters. Click Continue to Run in the Actions column. The system resumes the resharding task from where it was interrupted.
-
Cancel task
To cancel the task and restore the table to its original state, click Cancel Task in the Actions column. The system restores the table to its original state by removing the read-only state and deleting the temporary table.