OMS V3.3.1

Version information

• Version number: V3.3.1

• Previous version: V3.3.0

• Version release date: July 27, 2022

• Supported upgrade versions:

  • OMS V3.3.0 can be directly upgraded to V3.3.1.

  • Versions earlier than V3.2.1 are not supported. You must first upgrade them to V3.2.1. OMS V3.2.1 and later support only an upgrade to the immediately preceding version. The upgrade path is: V3.2.1 > V3.2.2 > V3.3.0 > V3.3.1.

Supported versions of data terminals

New features

Data migration and synchronization

• Added support for the ARM architecture.

• Added support for migrating data from a DB2 LUW database to the MySQL compatible mode of OceanBase Database.

• Added support for migrating data from a TiDB database to the MySQL compatible mode of OceanBase Database.

• Added support for migrating data from the MySQL compatible mode of OceanBase Database to a cobar database or sharding middleware. This feature also provides gray-scale switchover capabilities at the connection level.

• Extended support for migrating data from a MySQL 5.5 database to OceanBase Database.

• Added support for upgrading OceanBase Database V1.4.x to V2.2.x in a logical manner.

• Added estimation values for completed and remaining tasks in full migration and full verification projects, to help you estimate the actual completion time.

• Added performance data such as RPS, traffic, and baseline performance in full migration and full verification projects, to help you identify performance-related issues.

• Added the ability to import and export database objects, as well as row and column filter information, in CSV format.

• Defined the minimum privileges required for migrating data from an Oracle database to an OceanBase database. You can grant corresponding privileges to the migration user as needed.

Data source management

• Added new parameters, including configurl, drc_user, and drc_password, for physical OceanBase data sources. Added the __oceanbase_inner_drc_user username for migrating tables without primary keys in OceanBase Database. The main purpose is to resolve the dependency on OCP, reduce the deployment difficulty, and improve security.

• Added the address of the OceanBase proxy, database username, and password as new parameters for OceanBase logical data sources to adapt to more business scenarios.

Feature updates and optimizations

Interaction experience optimization

• Optimized the structure migration process to change unsupported or failed tables to be migrated after the structure migration is saved and executed. This solves the problem of low efficiency in repairing multiple tables.

• Added the functionality to refresh and view the latest logs for the component, to improve the efficiency of log viewing when a large number of logs are involved.

• Added topic-database-table as the display name for the target data source.

• Added error code and specific error information for component exceptions to improve the efficiency in troubleshooting.

• Added more prompts when migration or synchronization types are selected to inform users of relevant limitations in advance.

• Optimized the prompt language in the data migration project to help you understand the matching rules being used.

• Added a significant prompt when the project latency exceeds 300 seconds.

• Added the option to view the three quartiles of numeric data.

• Added explanations for the forward switchover steps.

Feature upgrades

• Optimized the latency display logic in OMS: X seconds (updated Y seconds ago). Normally, Y is less than 20. This helps you identify latency issues.

• In scenarios where the target table is an OceanBase partitioned table, full migration can achieve up to 500,000 RPS based on partition import, significantly reducing migration time.

• The incremental log component has been optimized to parse up to 5 TB of incremental log data per day, covering most customer scenarios.

• In scenarios where hot rows exist in the source, full migration time can be reduced based on hot row major compaction.

• Full verification now supports verifying only inconsistent table records, improving overall verification efficiency.

• Data verification now provides analysis of the reasons for inconsistencies and ROWID details (for the Oracle data source).

• When viewing parameter configurations for a data synchronization project, topic type and basic settings information are now displayed to facilitate viewing of project basics.

• Added the ability to manually enter topics for existing projects to avoid project creation issues caused by permission-related factors.

• Fixed some security issues.

Bug fixes

• Fixed the issue where filter statements without table names are not supported in the drop index syntax of incremental DDL operations.

• Fixed the issue where the incremental numeric format in the MySQL compatible mode of OceanBase Database is abnormal when data is synchronized to a Kafka data source, causing the connector to fail.

• Fixed the issue where, for a partitioned table as the source, checker sharding errors lead to read amplification during full migration of partitioned tables.

• Fixed the compatibility issues for upgrading OMS to V3.3.1 or V3.3.2 from V3.2.1 or V3.2.2.

  For OMS V3.2.1 or V3.2.2, you must add the dbType configuration item before upgrading. If you forget to add this configuration before the upgrade, the project will not run after the upgrade, and you will need to fix the issue.

  Notice

  If you are using OMS V3.3.0 but have data synchronization projects created in OMS V3.2.1 or V3.2.2, you may encounter the above issues. You can use the same method to fix the issue.

  The key to determining whether a fix is needed is to check if the Source config of the Connector component includes the dbType or db_type parameter. If either parameter is present, no fix is needed.

Known issues

• When you pause a full migration project, the project is displayed as paused, but the Checker component is still shown as running.

• During data synchronization from an Oracle database to a DataHub instance, the port of the Oracle store might be occupied.

• You cannot create a data source with the same name as an existing one created by the admin user in the OMS console if you log in as a non-admin user.

• After you run the stop oms_drc_supervisor command by using supervisorctl in the OMS Docker container, the corresponding process still exists.