Getting Error 1236 (Could not find first log file name in binary log index file) reading binlog in AWS DMS

Question

I have created two DMS task (full load + CDC) with RDS MYSQL 'Read' replica as source endpoint and REDSHIFT as target endpoint. For both the task source and target endpoint is same. One task with only 3 tables is running fine while another task with 400+ tables is failing with below error-

Error 1236 (Could not find first log file name in binary log index file) reading binlog in AWS DMS.

If I resume the task then it fails. However if I restart then it runs for a few days (6-7 days) and again it fails.

I have increased the log retention period to 24 hrs in both master and read replica of MySQL RDS instance but no luck.

Please assist me in resolving this issue. Thanks in advance.

Answer 1

AWS provides support documentation to help troubleshoot this error: https://aws.amazon.com/premiumsupport/knowledge-center/dms-cdc-error-1236-msql/

They show an example of increasing the retention to 24 hours, which I assume is why you tried that. Perhaps you need a greater retention period.

They also mention the possibility that the AWS DMS task got an error that caused the CDC to stop. In that case, it's possible that the CDC is offline for more than your retention period. If you don't resolve the cause of the CDC stopping, then no retention period will help, because if the CDC stops, then it will simply never catch up.

The support page describes a number of error scenarios that could cause the CDC to stop. If you need help, contact AWS Support. You're paying for support, so you might as well use it.