Manage migration jobs

Stop a running migration job at any time. During either the full dump phase or the CDC phase, the data migration is halted, and the migration job status changes to Stopping and subsequently to Stopped.

• If the migration job is stopped during the full dump phase, then you can delete or restart the job. However, WAL logs will continue to accumulate in the replication slots, eventually leading to your source database running out of disk space. In addition, Database Migration Service will keep pulling a replication delay, which means that Database Migration Service will continue to consume network traffic.
• If the migration job is stopped during the change data capture (CDC) phase, then you can promote or delete the job.

Monitor your database server's replication slot and disk size. You can choose to delete the migration job, which will clean all replication slots and network traffic.

If a migration was stopped during the CDC phase, then it can be resumed. Resuming the migration job resumes replication of the events that were accumulated during the stopped time.

If the migration failed and isn't recoverable, or was stopped during a full dump, then it can be restarted from the beginning.

If the replication has failed because of an error on one or more source or destination databases, then restarting the migration will wipe all data from any destination databases that are impacted by the errors.

If new databases were created on the source PostgreSQL server since starting the migration, then restarting the migration will also replicate the new databases into the destination instance.

• If the migration job status was Completed, then the record is deleted from the migration job list.
• If the migration job was in any other status, then there's an option to choose to delete the destination Cloud SQL instance/primary pair.