Open source
Schemaferry · Multi-Database Migration Orchestrator
A schema-aware orchestrator for zero-downtime migrations across PostgreSQL, MySQL, and MongoDB.
Zero-downtime migrations, validated by checksum
The problem
Moving data between databases is one of the few operations that can lose information silently. A cross-engine migration, say PostgreSQL to MongoDB, or a consolidation across MySQL and PostgreSQL, is usually a held-breath cutover: take the source down, run a script, hope the row counts match, and have no clean way back if they don’t. The risk scales with the dataset, and the verification is almost always an afterthought.
What I built
Schemaferry is a Python orchestrator that treats the migration as a verifiable, reversible operation rather than a one-shot script. It works across PostgreSQL, MySQL, and MongoDB, and handles the parts people skip: schema diffing to reconcile structural differences between engines, incremental sync so the cutover window stays small, data-validation checksums to prove the target matches the source, and rollback snapshots so a failed migration is an undo, not an incident.
The result
A migration becomes something you can prove succeeded and reverse if it didn’t. It has been used in production for zero-downtime migrations of large datasets, where the checksum validation is the difference between “the counts look right” and “every record is accounted for.” It is open source because safe migration tooling is something every team eventually needs and few have time to build.