Question

A retail company completed a heterogeneous migration of a 4 TB Oracle database to Amazon Aurora PostgreSQL using AWS DMS with continuous CDC replication. Before the final application cutover, the compliance team requires documented proof that every row in the source and target is identical, without pausing the ongoing CDC replication or introducing significant additional load on the source database. The team wants a repeatable, AWS-native approach to detect row-level and data-type mismatches. Which approach best meets these requirements?

Accepted Answer

Enable DMS data validation on the existing migration task so DMS compares source and target rows and reports mismatches in the validation statistics tables.

Answer

Write custom SQL scripts that compute MD5 checksums of each table on both databases and run them nightly from a scheduled Lambda function.

Answer

Stop the CDC task, take an Aurora snapshot, restore it to a separate cluster, and manually spot-check a sample of tables against the source.

Answer

Use AWS Glue to crawl both databases and run a Glue ETL job that joins the datasets to identify differences.