If a connector is deployed using a version that is prior to when log.mining.archive.log.only.mode was introduced and the offset SCN for the connector is still in the redo logs, starting the connector with this option will fail because the SCN validation check will determine that the SCN we're looking for isn't in the logs.

This check should be a bit more lenient. As for now a workaround, if enabling this feature, ahve the DBA flush all current redo logs to archive logs & restart the connector.