Per the documentation, a full binlog is required when using MySQL as a source. DMS will filter based on which databases and tables you are replicating.
based on the number of tables involved dms will query the binlog (log mining) for ddl/dml changes and wont get full binlog to dms instance.
There's no separation of tables in the binlog. All transactions / datapages for all tables (and objects) are kept there in serialized mode. Maybe in the future it could happen, given the source is open to handle this. This is done this way for consistency enforcement (transaction serialization) with rollback in mind either. Its very possible today to restore a database in a separate server, and then do the import on the required server Another alternative is to export the snapshot to s3 for the specific table. Check this out https://dplab.medium.com/restore-specific-table-data-from-rds-backup-eb3623786f5e
DMS task is creating full row duplicates in targetasked 5 months ago
Does AWS DMS load full binlog from source regardless if you only need a table or two?asked a month ago
Aurora Serverless MySQL to access the binlog or Activity Streamsasked 9 months ago
does AWS DMS support partition tables from postgresasked 8 months ago
Do source filters speed up DMS from AWS RDS to S3?asked 8 months ago
Does AWS DMS has any way to touch or modify source tables?asked 3 months ago
Can DMS do mapping that is not simply 1-1 from source to target?Accepted Answerasked 3 months ago
Why don't I get en event with binlog information when I create a snapshot?Accepted Answerasked a year ago
No binlog position from crash recovery is shown after restoring from RDS Aurora snapshotasked a month ago
DMS CDC - MySQL Binlog Precision - Seconds to Microsecondsasked 22 days ago