Migrate data from MySQL to PostgreSQL using ActiveRecord Based on https://github.com/braintree/mysql_to_postgresql

mysql_to_postgres.rb

#!/usr/bin/env ruby

require 'logger'
require 'active_record'
require 'active_record/base'
require 'active_support'

ActiveRecord::Base.logger = Logger.new("log/mysql_to_postgresql.log")

PROCESSES = 8
TABLES_WITHOUT_SEQUENCES = ["schema_migrations"]
RAILS_ENV = "development"
SKIP_TABLES = %w[]
TABLES = %w[]
db_config = YAML.load_file('config/database.yml')
MYSQL_CONNECTION_CONFIG = db_config["#{RAILS_ENV}_mysql"]
POSTGRES_CONNECTION_CONFIG = db_config["#{RAILS_ENV}_postgres"]

pids = []
PROCESSES.times do |process_index|
  pids << fork do
    class MysqlModelBase < ActiveRecord::Base
      establish_connection(MYSQL_CONNECTION_CONFIG)
    end

    class PostgresqlModelBase < ActiveRecord::Base
      establish_connection(POSTGRES_CONNECTION_CONFIG)
      def self.attributes_protected_by_default
        []
      end
    end

    tables = TABLES.empty? ? MysqlModelBase.connection.tables - SKIP_TABLES : TABLES

    tables.each_with_index do |table, table_index|
      next unless table_index % PROCESSES == process_index
      puts "[#{process_index}] Starting table: #{table}"

      id_column = case table
                when "schema_migrations"
                  "version"
                else
                  "id"
                end

      MysqlModelBase.table_name = table
      MysqlModelBase.reset_column_information
      PostgresqlModelBase.table_name = table
      PostgresqlModelBase.reset_column_information

      puts "[#{process_index}] -- truncate #{table}"
      PostgresqlModelBase.connection.execute %(TRUNCATE "#{table}" CASCADE;)

      index = 0
      MysqlModelBase.find_in_batches do |records|
        start = Time.now
        PostgresqlModelBase.transaction do
          records.each do |record|
            PostgresqlModelBase.create!(record.attributes)
          end
        end
        puts "[#{process_index}] -- #{table} - #{index} - #{records.size} per #{Time.now - start} sec"
        index += 1
      end

      # Set sequence values
      unless TABLES_WITHOUT_SEQUENCES.include?(table)
        PostgresqlModelBase.connection.execute("select setval('#{table}_#{id_column}_seq', (select max(#{id_column}) from #{table}))")
      end
    end
  end
end
pids.each { |pid| Process.waitpid(pid) }