What's the impact of data sync on the target database? Does the target database need to be empty?
Syncing data with DTS won't impact the target database, and the target database doesn't need to be empty. DTS will check whether there are objects with the same name in the source and target databases. If there is a duplicate name conflict, you can handle it by using the configured conflict resolution policy. You will be prompted of an error, and you can ignore it.
Will the sync service be affected if high availability (HA) switch occurs between the source and target instances?
If the source instance supports and enables global transaction identifier (GTID), when HA switch occurs in the source instance during incremental sync, the service will be automatically reconnected, and the synced data flow will resume soon after HA switch is completed.
If HA switch occurs in the target instance during incremental sync, the service will also be automatically reconnected, and the synced data flow will resume soon after HA switch is completed.
Can data in an instance on a later version be synced to an instance on an earlier version?
No. For data sync between instances of the same type, the major version of the target instance must be later than that of the source instance. Taking MySQL as an example, data cannot be synced from MySQL 5.7 to MySQL 5.6.
Can the source/target instance be a local instance?
Yes. Data can be synced from TencentDB to a local database over the public network and CCN.
Can I perform two-way sync for DDL statements in a two-way sync topology?
No. During sync instance creation, DDL statements in only one instance can be synced; otherwise, the algorithm will detect a DDL loop and then prohibit the creation of one of the instances.
Are non-transactional engines supported?
The current technical solution adds routing information to transactions to mark transaction sources, which relies on the atomicity of transactions. However, databases/tables based on non-transactional engines will corrupt such atomicity, and data consistency cannot be guaranteed. Therefore, we recommend that you not use non-transactional engines.
Can I add or delete sync objects after a sync task is started?
Yes. You can modify the sync task configurations by selecting the target sync task and clicking Modify Sync Configuration in the Operation column.
On the configuration modification page, you can add/delete sync objects and modify the primary key conflict policy or SQL sync policy. When you modify the sync configuration, the running sync task won't be paused or affected. For more information, see Modifying Sync Configuration.