Database schema migration is the systematic evolution of database structures alongside application code — a critical practice for maintaining data integrity, enabling continuous delivery, and supporting team collaboration in modern software development. Unlike application code, database changes are inherently stateful and often irreversible, requiring careful planning to avoid downtime, data loss, or production incidents. Whether managing incremental schema changes with migration-based tools like Flyway and Liquibase, orchestrating zero-downtime transitions through expand-and-contract patterns, or coordinating parallel development across multiple team branches, the migration strategy you choose directly impacts deployment risk, development velocity, and system reliability. This cheat sheet covers versioned migration workflows, zero-downtime techniques, cross-database tooling, and team collaboration patterns that experienced teams rely on to ship schema changes safely and confidently.
What This Cheat Sheet Covers
This topic spans 11 focused tables and 53 indexed concepts. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.
Table 1: Popular Migration Tools
| Tool | Example | Description |
|---|---|---|
flyway migrate | • Migration-first approach using versioned SQL scripts prefixed with V1__, V2__ — applies sequential migrations tracked in flyway_schema_history table.• Supports Java, Python, Node.js via CLI or embedded library • No-code-generation philosophy simplifies auditing and debugging | |
<changeSet id="1" author="dev"> <createTable tableName="users"/></changeSet> | • XML/YAML/JSON changelogs define database-agnostic migrations with built-in rollback generation. • Strong enterprise governance with preconditions and contexts • State-based comparisons available via commercial editions | |
gh-ost --host=prod --alter="ADD COLUMN status VARCHAR(20)" | • GitHub's triggerless online schema migration for MySQL — reads binary log instead of triggers to avoid overhead. • Throttles automatically based on replication lag • Allows pause/resume & cutover control during migration | |
pt-osc --alter "ADD INDEX idx_email (email)" D=mydb,t=users | • Percona's trigger-based online migration for MySQL — creates shadow table, copies data via chunking, uses triggers to sync writes. • Well-tested in large-scale production environments • More overhead than gh-ost due to triggers |