MySQL

Do not use Finch in production. It’s a development tool intended for use only in insecure development environments.

Finch defaults to username finch, password amazing , host 127.0.0.1 (via TCP).

Finch is a development tool intended for use only in insecure development environments, but a safer MySQL user is suggested:

CREATE USER finch@'%' IDENTIFIED BY 'amazing';

GRANT SELECT, INSERT, UPDATE, DELETE, CREATE, DROP,  REFERENCES, INDEX, ALTER,  CREATE TEMPORARY TABLES, LOCK TABLES
ON finch.* TO finch@'%';

If you want to use your local root MySQL user, setting environment variable FINCH_DSN is the easiest way to override the defaults.

Finch checks stage.mysql.dsn first: if set, it’s used and processing stops. --dsn sets and overwrites stage.mysql.dsn from the command line: stage.mysql.dsn = --dsn. If stage.mysql.dsn is not set, it inherits _all.yaml mysql.dsn.

If no DSN is specified, Finch loads the other stage.mysql settings, which inherit values from _all.yaml mysql.

There are three ways to set the default database:

The connection default database is standard: the database that the client requires when connecting to MySQL. It’s specified by --dsn, or by --database, or by mysql.db. If the database doesn’t exist, the connection fails with an error:

% mysql -D does_NOT_exist
ERROR 1049 (42000): Unknown database 'does_not_exist'

The connection database is useful to ensure that all clients use it, but there’s a problem: it doesn’t work if the stage is going to create the database.

It is not recommended to have a stage drop schemas or tables in a trx file. A human should do this manually to ensure only the correct schemas or tables are dropped.

Although not recommended, a stage can create its own schema in a trx file like:

DROP DATABASE IF EXISTS finch;

CREATE DATABASE finch;

USE finch;

CREATE TABLE ...

That’s an example of a trx default database: using an explicit USE db statement, which Finch allows.

The client group default database is a special case: it make Finch execute USE db once for each client in the group during stage preparation (before the stage runs), which avoids having to use a trx default database. This is done after connecting, so the database doesn’t need to exist to connect, but it needs to exist when the client group is prepared (else --test will fail).