Data Transmission Service:Upgrade a change tracking instance from the previous version to the new version

Usage notes

• Change tracking instances of the previous version and change tracking instances of the new version have different architectures and are implemented in different ways. Therefore, you need to create a change tracking instance of the new version to consume tracked data.
• Then, you can release your change tracking instance of the previous version. A released instance cannot be restored. Make sure that the change tracking instance of the new version can consume tracked data before you release the change tracking instance of the previous version.

Procedure

1. Create a change tracking instance of the new version. For more information, see Overview of change tracking scenarios.
2. Consume tracked data. For more information, see Consume tracked data.

   Table 1. Comparison between change tracking instances of the previous and new versions

   Item	Change tracking instance of the previous version	Change tracking instance of the new version
   Consumption mode	An AccessKey pair is used to consume data.	A consumer group is used to consume data. For more information, see Create consumer groups.
   Consumption checkpoint management	A change tracking instance uses only one consumption checkpoint. You can edit the consumption checkpoint of a change tracking task in the DTS console.	Each consumer group uses a separate consumer offset. Change tracking instances of the new version support flexible and diverse consumption checkpoint management. If an SDK client works in SUBSCRIBE mode, the consumption checkpoint saved on the DTS server is used when the SDK client is restarted. In this mode, the consumption checkpoint cannot be modified. If an SDK client works in ASSIGN mode, the specified consumption checkpoint is used when the SDK client is restarted. For more information, see Save and query the consumption checkpoint.
   SDK used by the client	You must use the SDK of the previous version to consume tracked data.	You can use the SDK of the new version to consume tracked data. This method is recommended. For more information, see Use the SDK demo to consume tracked data. You can use a Flink client to consume tracked data. For more information, see Use flink-dts-connector to consume tracked data. You can use a native Kafka client to consume tracked data. For more information, see Use a Kafka client to consume tracked data.
   Message format	The format of the previous version is used.	Change tracking SDK of the new version: DefaultUserRecord Flink client: DataStream API, Table API, and SQL Kafka client: Avro
   Troubleshooting	N/A	Troubleshoot change tracking issues based on errors reported by the SDK of the new version: Troubleshooting Troubleshoot change tracking issues by using a dedicated DTS toolkit: Troubleshoot issues with change tracking tasks

3. After the change tracking instance of the new version can consume the data at a specific consumption checkpoint, you can release the change tracking instance of the previous version. For more information, see Release DTS instances.