Backup Stages

When mydumper start to take a backup needs to execute multiple task and reach milestone.

Determine backup strategy

There are multiple locking options that you can instruct mydumper to take backups. You should review which locking mechanisim is the best for you.

Send locking statements

mydumper establishes a primary main connection and in some cases a secondary main connection. We need them to synchronize the working threads.

Start jobs creation

At this point we are able to start to create jobs depending the filters that has been used. For instance, if you used -B, we need to export the schema definition but we also need to create the jobs to export the schema of the tables and the data in it.

Start the dump threads

As the locking mechanisim is in please we are allow to create the working threads to start processing the jobs.

Release lockings

Once the working threads are in sync, the main connection is allowed to release lockings related to the synchronization of the workers.

Wait backup to complete

The workers will be processing the jobs, exporting the schema and data. The main thread will be waiting until they finish.

Releasing remaining lockings

Depending on the locking mechanisim that you configure, we might need to release the remaining locks.