pg2mysql

pg2mysql was created to facilitate migrating data from PostgreSQL to MySQL given mostly equivalent schemas.

In PostgreSQL it is common to use the text datatype for character data, which the PostgreSQL documentation describes as having effectively no limit. In MySQL this is not the case, as the datatype with the same name (text) is limited to 65535, and the more common datatype, varchar, is defined with an explicit limit (e.g. varchar(255)).

This means that, given a column with text datatype in PostgreSQL, there must be enough room in the equivalent MySQL column for the data in PostgreSQL to be safely migrated over. This tool can be used to validate the target MySQL schema against a populated PostgreSQL database and, provided the data in the PostgreSQL database is compatible, the migration to move the data from PostgreSQL to MySQL.

Install from source

go get github.com/pivotal-cf/pg2mysql/cmd/pg2mysql

Usage

Create a config:

$ cat > config.yml <<EOF
mysql:
  database: some-dbname
  username: some-user
  password: some-password
  host: 192.168.10.1
  port: 3306

postgresql:
  database: some-dbname
  username: some-user
  password: some-password
  host: 192.168.10.2
  port: 5432
  ssl_mode: disable
EOF

Note: See PostgreSQL documentation for valid SSL mode values.

Run the validator:

$ pg2mysql -c config.yml validate
found incompatible rows in apps with IDs [2]
found incompatible rows in app_usage_events with IDs [9 10 11 12]
found incompatible rows in events with IDs [16 17 18]

If there are any incompatible rows, as in above, they will need to be modified before proceeding with a migration.

Run the migrator:

$ pg2mysql -c config.yml migrate --truncate
inserted 1 records into spaces_developers
inserted 0 records into security_groups_spaces
inserted 0 records into service_bindings
inserted 2 records into droplets
inserted 2 records into organizations
inserted 3 records into lockings
inserted 0 records into service_dashboard_clients
inserted 0 records into route_bindings
...

Note: The --truncate flag will truncate each table prior to copying data over.