XUANWU_V2 table engine features and compaction service-AnalyticDB(AnalyticDB)-阿里云帮助中心

Prerequisites

An AnalyticDB for MySQL Enterprise Edition, Basic Edition, or Data Lakehouse Edition cluster is created.
The minor version of the cluster is 3.2.2.0 or later.

Note
To view and update the minor version, go to the Configuration Information section on the Cluster Information page in the AnalyticDB for MySQL console.

Key features

XUANWU_V2 enables AnalyticDB for MySQL to deliver high-throughput real-time writes and queries with these key features:

Storage medium: Data is stored in OSS with a cloud disk cache, reducing storage costs while maintaining query performance and enabling faster horizontal scaling and higher elasticity.
Storage format: A next-generation columnar format provides fine-grained control over memory and disk I/O, delivering higher I/O concurrency with lower memory consumption and reduced GC impact.
Independent compaction component: Compaction is offloaded to an independent service, improving online query and write stability while providing higher compaction throughput and flexible resource scheduling.

Limitations

The following table compares feature support between the XUANWU and XUANWU_V2 engines.

Feature	XUANWU engine	XUANWU_V2 engine
regular index	By default, a full-column index is created (`INDEX_ALL='Y'`).	By default, XUANWU_V2 creates an index only for the primary key. No indexes are created for other columns (`INDEX_ALL='N'`). Note You can create or modify a full-column index in one of the following ways: Before you create a table: Run the `SET ADB_CONFIG XUANWUV2_DEFAULT_INDEX_ALL=true` statement at the cluster level. This creates a full-column index for the current table and all subsequent new tables. When you create a table: Explicitly specify `INDEX_ALL='Y'` in the CREATE TABLE statement. This creates a full-column index for the current table. After you create a table: Run an ALTER TABLE statement to modify the full-column index of an existing table.
vector index	Supported	Not supported
binary log	Supported	Not supported
Spark elastic import	Supported	Not supported

You cannot switch between the XUANWU and XUANWU_V2 engines. To switch, create a new table with the desired engine and migrate the data.

For example, to switch a test table from XUANWU_V2 to XUANWU, create a new XUANWU table and migrate the data from test.

Billing

By default, data in XUANWU_V2 tables is hot data and is billed based on the hot data storage space. If you enable tiered storage of hot and cold data, you are charged for both hot and cold data storage. Pricing for Enterprise Edition and Basic Edition and Pricing for Data Lakehouse Edition.

Note

Log on to the AnalyticDB for MySQL console and go to the Storage Overview page to view XUANWU_V2 table storage. View the data size of a table.

Configure the table engine

For clusters with minor version 3.2.2.0 or later, XUANWU_V2 is the default table engine. You can override this at the cluster or table level.

Cluster-level configuration

Run the following command to set the default engine for the entire cluster:

SET ADB_CONFIG RC_DDL_ENGINE_REWRITE_XUANWUV2=false|true;

Valid values:

true (default): Existing tables are unchanged. New tables use XUANWU_V2 even if you specify ENGINE=XUANWU.

Note
To create a XUANWU table, add the /* RC_DDL_ENGINE_REWRITE_XUANWUV2=false */ hint before CREATE TABLE.
false: Existing tables are unchanged. New tables default to XUANWU. To create a XUANWU_V2 table, specify ENGINE=XUANWU_V2 or add the /* RC_DDL_ENGINE_REWRITE_XUANWUV2=true */ hint before CREATE TABLE.

Query the current configuration:

SHOW ADB_CONFIG KEY=RC_DDL_ENGINE_REWRITE_XUANWUV2;

Table-level configuration

Important

Table-level configurations are supported only on clusters with minor version 3.2.2.12 or later (excluding minor versions 3.2.3.1 and 3.2.3.2).

A table-level hint applies only to the current CREATE TABLE statement and overrides the cluster-level setting.

/* RC_DDL_ENGINE_REWRITE_XUANWUV2=false|true */

Valid values:

true: Existing tables are unchanged. The table created by this statement uses XUANWU_V2 even if you specify ENGINE=XUANWU.
false: Existing tables are unchanged. The table created by this statement uses XUANWU by default. To use XUANWU_V2, specify ENGINE=XUANWU_V2.

Enable the disk cache

Warning

After you enable the disk cache, the cluster automatically restarts. This may cause transient disconnections. We recommend that you perform this operation during off-peak hours and ensure that your application has an automatic reconnection mechanism.

XUANWU_V2 cold data is stored in OSS. The disk cache improves random read performance for cold data. You are charged for the configured cache space.

Do not disable the disk cache after enabling it. Doing so may cause severe performance degradation.

Log on to the AnalyticDB for MySQL console. In the upper-left corner of the console, select a region. In the left-side navigation pane, click Clusters. Find the cluster that you want to manage and click the cluster ID.
On the Cluster Information page, in the Configuration Information section, click Configure next to Cluster Query Acceleration Configuration.
On the Disk Cache tab, click the switch and set the cache size.

Note

After you enable the Disk Cache, you can also follow the preceding steps to view the cache size.

Enable the compaction service

The compaction service offloads compaction to an independent resource pool, reducing online resource consumption and improving stability. This optional feature is disabled by default, billed pay-as-you-go, and can be disabled at any time. Pricing for Enterprise Edition and Basic Edition and Pricing for Data Lakehouse Edition.

Recommended for services with:

High online CPU and memory utilization.
Performance significantly affected by periodic compaction.

Procedure

Log on to the AnalyticDB for MySQL console. In the upper-left corner of the console, select a region. In the left-side navigation pane, click Clusters. Find the cluster that you want to manage and click the cluster ID.
On the Cluster Information page, in the Configuration Information section, click Configure next to Cluster Query Acceleration Configuration.
On the Remote Build tab, click the switch to enable the service.