How to recover from invalid resume token on DMS MongoDB connector?

0

When we pause our mongo connections on DMS for some hours and the resume point is no longer in the oplog we are unable to restart or resume our task.

Is there any way we can ignore these errors and continue replication or setup a hearbeat similar to Postgres RDS endpoint to update the resume point? The only fix we found was deleting the task and recreating it.

Resume of change stream was not possible, as the resume point may no longer be in the oplog
asked 2 years ago1245 views
1 Answer
0

Above error basically means the last oplog position when the dms task was stopped is no longer and there are chances of data loss due to missing transactions. I would suggest increasing the oplog so that the required entry postion is not overwritten

https://www.mongodb.com/docs/manual/tutorial/change-oplog-size/#c.-change-the-oplog-size-of-the-replica-set-member

AWS
answered 2 years ago
  • How can I recover after the problem has already happened withou deleting the task? Our main issue is because our task are created with terraform. We need to open a PR to fix this issue by recreating the resource because actions available on the UI won't recover from the error. Kafka mongo connector has an entire topic on how to recover from the error but I couldn't find anything similar for Mongo DMS.

    https://www.mongodb.com/docs/kafka-connector/current/troubleshooting/recover-from-invalid-resume-token/

    It looks like DMS never updates the "Change data capture (CDC) recovery checkpoint". We are unable to keep all the history in the oplog forever. How can we change the CDC to update the recovery checkpoint or restart it programatically?

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions