mmk redo

Rolls a migration back and runs it again, in one command. Without a file name it uses the migration that ran most recently.

npx mmk redo [file] [options]

Examples

npx mmk redo           # the most recently applied migration
npx mmk redo <file>    # a specific one

When to use it

Mostly while you're writing a migration. You run it, look at the data, change your mind, edit the file, and want to try again from a clean state:

$ npx mmk redo
Loaded config from mmk.config.js
↩ Reverted 20261015224137-index-orders-by-customer.js   [5ms]
✔ Applied  20261015224137-index-orders-by-customer.js   [128ms]

Options

Option	What it does
[file]	Redo this file instead of the most recent one.
--no-lock	Don't take the lock. Only for local development.
--json	Print the results as JSON on stdout.

The global flags work here too.

Good to know

• One lock for both halves. The rollback and the re-run happen under a single lock, so nothing else can run in between, and the migration can't be left rolled back if the second half fails to start.
• Edited files are fine. Unlike mmk down, redo doesn't refuse a file that changed since it ran. Editing it is usually why you're running redo. It runs the current down(), so make sure that still matches what the old up() did.
• Both events are recorded. The history shows the rollback and the new run.
• Imported migrate-mongo records can't be redone, for the same reason they can't be rolled back.

Exit codes

Code	Meaning
0	Rolled back and applied again.
1	Either half failed, the file isn't applied, or it was imported from migrate-mongo.