= ActiveRecord::Turntable {}[http://travis-ci.org/drecom/activerecord-turntable] ActiveRecord::Turntable is a database sharding extension for ActiveRecord. == Dependencies activerecord(>=3.0.0) == Supported Database Currently supports mysql only. == Installation Add 'activerecord-turntable' to Gemfile: gem 'activerecord-turntable', '1.0.1' Run a bundle install: bundle install Run install generator: bundle exec rails g active_record:turntable:install generator creates Rails.root/config/turntable.yml == ActiveRecord::Turntable terms * Shard * Cluster - a cluster of shards. * Master - default ActiveRecord::Base's connection * Sequencer - a module that creates global sequence number == Example === Example Databases Structure one main database(default ActiveRecord::Base connection) and three user databases sharded by user_id. +-------+ | App | +-------+ | +---------+-----------+---------+---------+ | | | | | +--------+ +---------+ +-------+ +-------+ +-------+ | Master | |Sequencer| |UserDB1| |UserDB2| |UserDB3| +--------+ +---------+ +-------+ +-------+ +-------+ === Example Configuration Edit turntable.yml and database.yml. See below example config. example turntable.yml: development: clusters: user_cluster: # <-- cluster name algorithm: range # <-- only range or range_bsearch sharding is supported currently. seq: connection: user_seq_1 shards: - connection: user_shard_1 # <-- shard name less_than: 100 # <-- shard range(like mysql partitioning) - connection: user_shard_2 less_than: 200 - connection: user_shard_3 less_than: 2000000000 database.yml: connection_spec: &spec adapter: mysql2 encoding: utf8 reconnect: false pool: 5 username: root password: root socket: /tmp/mysql.sock development: <<: *spec database: sample_app_development seq: # <-- sequence definition user_seq_1: <<: *spec database: sample_app_user_seq_development shards: # <-- shards definition user_shard_1: <<: *spec database: sample_app_user1_development user_shard_2: <<: *spec database: sample_app_user2_development user_shard_3: <<: *spec database: sample_app_user3_development === Migration Example Generate a model: bundle exec rails g model user name:string Edit migration file: class CreateUsers < ActiveRecord::Migration clusters :user_cluster # <-- the cluster that executes migration def change create_table :users do |t| t.string :name t.timestamps end create_sequence_for(:users) end end then please execute rake tasks: bundle exec rake db:create bundle exec rake db:migrate Those rake tasks would be executed to shards too. === Model Example add turntable [shard_key_name] to the model class: class User < ActiveRecord::Base turntable :user_cluster, :id sequencer has_one :status end class Status < ActiveRecord::Base turntable :user_cluster, :user_id sequencer belongs_to :user end == Usage === Create > User.create(name: "hoge") (0.0ms) [Shard: user_seq_1] BEGIN (0.3ms) [Shard: user_seq_1] UPDATE `users_id_seq` SET id=LAST_INSERT_ID(id+1) (0.8ms) [Shard: user_seq_1] COMMIT (0.1ms) [Shard: user_seq_1] SELECT LAST_INSERT_ID() (0.1ms) [Shard: user_shard_1] BEGIN [ActiveRecord::Turntable] Sending method: insert, sql: #, shards: ["user_shard_1"] SQL (0.8ms) [Shard: user_shard_1] INSERT INTO `users` (`created_at`, `id`, `name`, `updated_at`) VALUES ('2012-04-10 03:59:42', 2, 'hoge', '2012-04-10 03:59:42') (0.4ms) [Shard: user_shard_1] COMMIT => # === Retrieve > user = User.find(2) [ActiveRecord::Turntable] Sending method: select_all, sql: #, shards: ["user_shard_1"] User Load (0.3ms) [Shard: user_shard_1] SELECT `users`.* FROM `users` WHERE `users`.`id` = 2 LIMIT 1 => # === Update > user.update_attributes(name: "hogefoo") (0.1ms) [Shard: user_shard_1] BEGIN [ActiveRecord::Turntable] Sending method: update, sql: UPDATE `users` SET `name` = 'hogefoo', `updated_at` = '2012-04-10 04:07:52' WHERE `users`.`id` = 2, shards: ["user_shard_1"] (0.3ms) [Shard: user_shard_1] UPDATE `users` SET `name` = 'hogefoo', `updated_at` = '2012-04-10 04:07:52' WHERE `users`.`id` = 2 (0.8ms) [Shard: user_shard_1] COMMIT => true === Delete > user.destroy (0.2ms) [Shard: user_shard_1] BEGIN [ActiveRecord::Turntable] Sending method: delete, sql: #, shards: ["user_shard_1"] SQL (0.3ms) [Shard: user_shard_1] DELETE FROM `users` WHERE `users`.`id` = 2 (1.7ms) [Shard: user_shard_1] COMMIT => # === Count > User.count [ActiveRecord::Turntable] Sending method: select_value, sql: #, shards: ["user_shard_1", "user_shard_2", "user_shard_3"] (0.8ms) [Shard: user_shard_1] SELECT COUNT(*) FROM `users` (0.3ms) [Shard: user_shard_2] SELECT COUNT(*) FROM `users` (0.2ms) [Shard: user_shard_3] SELECT COUNT(*) FROM `users` => 1 === Sequencer Sequencer provides generating global IDs. Add below to the migration: create_sequence_for(:users) # <-- this line creates sequence table This will creates sequence table. Next, add sequencer to the model: class User < ActiveRecord::Base turntable :id sequencer # <-- this line enables sequencer module has_one :status end === Transaction > user = User.find(2) > user3 = User.create(name: "hoge3") > User.shards_transaction([user, user3]) do > user.name = "hogehoge" > user3.name = "hogehoge3" > user.save! > user3.save! > end === Send query to a specific shard. use with_shard method: AR::Base.connection.with_shard(shard1) do # something queries to shard1 end To access shard objects, use below: * AR::Base.connection.shards # \\{shard_name => shard_obj,....} * AR::Base#turntable_shard * AR::Base.connection.select_shard(shard_key_value) #=> shard === Send query to all shards use with_all method: User.connection.with_all do User.order("created_at DESC").limit(3).all end === Migration If you specify cluster or shard, migration will be executed to the cluster(or shard) and master database. to specify cluster: class CreateUsers < ActiveRecord::Migration clusters :user_cluster .... end to specify shard: class CreateUsers < ActiveRecord::Migration shards :user_shard_01 .... end if you would like to execute a migration to all databases in the environments, use :all symbol: class CreateUsers < ActiveRecord::Migration clusters :all .... end == Connection Management Rails's ConnectionManagement middleware keeps ActiveRecord's connection during the process is alive, but Turntable keeps more connections. This may cause flooding max connections on your database. So, we made a middleware that disconnects on each request. To use turntable's ConnectionManagement middleware, add below line to your initializer. app.middleware.swap ActiveRecord::ConnectionAdapters::ConnectionManagement, ActiveRecord::Turntable::Rack::ConnectionManagement == Inability * Queries includes "ORDER BY", "GROUP BY" and "LIMIT" clauses cannot be distributed. * "has many through" and "habtm" relationships causes wrong results. ex) User-Friend-User relation == Thanks to ConnectionProxy, Distributed Migration implementation is inspired by Octopus and DataFabric. == License activerecord-turntable is released under the MIT license: Copyright (c) 2012 Drecom Co.,Ltd. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.