Hi all, our team is currently using DMS to capture CDC data from our source RDS DB to S3 target. Occasionally, we see that DMS is writing multiple updates for the same record with a given ID to an S3 file with the same CDC timestamp (set by enabling the timestamp_column_name
setting). Some questions around this:
- Can I rely on the order of the record stored in the file to represent the transaction order?
- I read up on the doc that DMS has
preserve_transaction
S3 target setting, keen to know this a bit further as I am currently using date_partition_enabled
setting which cannot be used in tandem with preserve_transaction
Cheers