While migrations are sufficient for most needs, there are certain situations that require you to execute the same action over and over again. This could be recompiling procedures, updating materialized views and many other types of housekeeping.
For this reason, Flyway offers you the possibility to hook into its lifecycle by using Callbacks.
These are the events Flyway supports:
|beforeMigrate||Before Migrate runs|
|beforeRepeatables||Before all repeatable migrations during Migrate|
|beforeEachMigrate||Before every single migration during Migrate|
|beforeEachMigrateStatement Flyway Teams||Before every single statement of a migration during Migrate|
|afterEachMigrateStatement Flyway Teams||After every single successful statement of a migration during Migrate|
|afterEachMigrateStatementError Flyway Teams||After every single failed statement of a migration during Migrate|
|afterEachMigrate||After every single successful migration during Migrate|
|afterEachMigrateError||After every single failed migration during Migrate|
|afterMigrate||After successful Migrate runs|
|afterMigrateApplied||After successful Migrate runs where at least one migration has been applied|
|afterVersioned||After all versioned migrations during Migrate|
|afterMigrateError||After failed Migrate runs|
|beforeUndo Flyway Teams||Before Undo runs|
|beforeEachUndo Flyway Teams||Before every single migration during Undo|
|beforeEachUndoStatement Flyway Teams||Before every single statement of a migration during Undo|
|afterEachUndoStatement Flyway Teams||After every single successful statement of a migration during Undo|
|afterEachUndoStatementError Flyway Teams||After every single failed statement of a migration during Undo|
|afterEachUndo Flyway Teams||After every single successful migration during Undo|
|afterEachUndoError Flyway Teams||After every single failed migration during Undo|
|afterUndo Flyway Teams||After successful Undo runs|
|afterUndoError Flyway Teams||After failed Undo runs|
|beforeClean||Before Clean runs|
|afterClean||After successful Clean runs|
|afterCleanError||After failed Clean runs|
|beforeInfo||Before Info runs|
|afterInfo||After successful Info runs|
|afterInfoError||After failed Info runs|
|beforeValidate||Before Validate runs|
|afterValidate||After successful Validate runs|
|afterValidateError||After failed Validate runs|
|beforeBaseline||Before Baseline runs|
|afterBaseline||After successful Baseline runs|
|afterBaselineError||After failed Baseline runs|
|beforeRepair||Before Repair runs|
|afterRepair||After successful Repair runs|
|afterRepairError||After failed Repair runs|
|createSchema||Before automatically creating non-existent schemas|
|beforeConnect Flyway Teams||Before Flyway connects to the database|
Callbacks can be implemented either in SQL or in Java.
The most convenient way to hook into Flyway’s lifecycle is through SQL callbacks. These are simply sql files in the configured locations following a certain naming convention: the event name followed by the SQL migration suffix.
Using the default settings, Flyway looks in its default locations (<install_dir>/sql) for the Command-line tool)
for SQL files like
Placeholder replacement works just like it does for SQL migrations.
Optionally the callback may also include a description. In that case the callback name is composed of the event name,
the separator, the description and the suffix. Example:
Note: Flyway will also honor any
sqlMigrationSuffixes you have configured, when scanning for SQL callbacks.
If SQL Callbacks aren’t flexible enough for you, you have the option to implement the Callback interface yourself. You can even hook multiple Callback implementations in the lifecycle. Java callbacks have the additional flexibility that a single Callback implementation can handle multiple lifecycle events, and are therefore not bound by the SQL callback naming convention.
More info: Java-based Callbacks
Much like SQL callbacks, Flyway also supports the execution of callbacks written in a scripting language. The supported file extensions are the same as those supported by script migrations. For example, you could have a
beforeRepair__vacuum.ps1 callback. Script callbacks give you much more flexibility and power during the migration lifecycle. Some of the things you can achieve are:
When multiple callbacks for the same event are found, they are executed in the alphabetical order.
Click here to see a tutorial on using callbacks.