mmk redo

Roll back then re-apply a migration in one step — the last applied, or a specific file.

mmk redo [file] [options]

Usage

mmk redo            # down + up the most recently applied migration
mmk redo <file>     # down + up a specific migration

How it works

mmk redo runs down() then up() for the target. It's the fast loop for iterating on a migration during development:

↩ Reverted 20260605120000-add-users-index.js     [18ms]
✔ Applied  20260605120000-add-users-index.js     [40ms]

Options

Option	Description
[file]	Redo a specific migration file instead of the last applied one.
--no-lock	Skip the concurrency lock. Dev only.
--json	Emit the run results as a JSON array on stdout.

Plus the global flags.

Notes

• The whole down → up runs under a single lock acquisition, so no other process can interleave between the revert and the re-apply (and the migration can't be stranded in a reverted state if the re-apply fails to start).
• The down half is forced past the checksum guard — redoing an edited migration is the usual reason to run redo, so a drifted file does not block it.
• redo inherits the forward-only guard: an imported migrate-mongo record cannot be redone and is rejected up front.
• Because the down half reverts the record and the up half re-applies it, the audit trail keeps both events.

Exit codes

Code	Meaning
0	The migration was reverted and re-applied.
1	Either half threw, or the target isn't applied / is irreversible.