For beginners¶

mydumper¶

You need to use --outputdir to set the directory of the backup or a backup dir will be created.

What do you want to backup? a single database? multiple databases? a few tables?¶

You can use

-B with a list of databases
-T with a list of tables
multiple instances of --regex are allowed

What kind of lockings can be used?¶

mydumper needs to sync the threads and it provides different options based on your capabilities https://mydumper.github.io/mydumper/docs/html/locks.html#synchronizing-threads

What about DDL locks?¶

https://mydumper.github.io/mydumper/docs/html/locks.html#ddl-locks

Do you have non transactional tables?¶

If you do, you must set --trx-tables=0 or --no-trx-tables (if available)

Do you want to compress the backup?¶

fyi: zstd takes the same amount of time than raw files and reduce it like gzip compression

https://mydumper.github.io/mydumper/docs/html/exec_per_thread.html

Do you want a single backup file?¶

Do you need the replica position?¶

Use --replica-data

Do you want to check the consistency?¶

Use Use -M

If you are not sure follow this guidance:

if it is a migration, you should do it
if it is just for testings, you need it fast and it is a large backup, you can avoid it.
if it is a daily backup you can avoid it to reduce the backup timings
if it is a daily backup that you are going to test, then you must use it

myloader¶

Are you going to replace the tables? -o/--drop-tables?¶

Use --drop-tables