Module: ActiveRecord::ConnectionHandling

Included in:
Base
Defined in:
activerecord/lib/active_record/connection_handling.rb

Overview

Active Record Connection Handling

Constant Summary collapse

RAILS_ENV =
-> { (Rails.env if defined?(Rails.env)) || ENV["RAILS_ENV"].presence || ENV["RACK_ENV"].presence }
DEFAULT_ENV =
-> { RAILS_ENV.call || "default_env" }

Instance Attribute Summary collapse

Instance Method Summary collapse

Instance Attribute Details

#connection_specification_nameObject

Returns the connection specification name from the current class or its parent.



316
317
318
319
320
321
# File 'activerecord/lib/active_record/connection_handling.rb', line 316

def connection_specification_name
  if @connection_specification_name.nil?
    return self == Base ? Base.name : superclass.connection_specification_name
  end
  @connection_specification_name
end

Instance Method Details

#adapter_classObject

:nodoc:



338
339
340
# File 'activerecord/lib/active_record/connection_handling.rb', line 338

def adapter_class # :nodoc:
  connection_pool.db_config.adapter_class
end

#clear_cache!Object

:nodoc:



372
373
374
# File 'activerecord/lib/active_record/connection_handling.rb', line 372

def clear_cache! # :nodoc:
  connection_pool.schema_cache.clear!
end

#clear_query_caches_for_current_threadObject

Clears the query cache for all connections associated with the current thread.



258
259
260
261
262
# File 'activerecord/lib/active_record/connection_handling.rb', line 258

def clear_query_caches_for_current_thread
  connection_handler.each_connection_pool do |pool|
    pool.clear_query_cache
  end
end

#connected?Boolean

Returns true if Active Record is connected.

Returns:

  • (Boolean)


351
352
353
# File 'activerecord/lib/active_record/connection_handling.rb', line 351

def connected?
  connection_handler.connected?(connection_specification_name, role: current_role, shard: current_shard)
end

#connected_to(role: nil, shard: nil, prevent_writes: false, &blk) ⇒ Object

Connects to a role (e.g. writing, reading, or a custom role) and/or shard for the duration of the block. At the end of the block the connection will be returned to the original role / shard.

If only a role is passed, Active Record will look up the connection based on the requested role. If a non-established role is requested an ActiveRecord::ConnectionNotEstablished error will be raised:

ActiveRecord::Base.connected_to(role: :writing) do
  Dog.create! # creates dog using dog writing connection
end

ActiveRecord::Base.connected_to(role: :reading) do
  Dog.create! # throws exception because we're on a replica
end

When swapping to a shard, the role must be passed as well. If a non-existent shard is passed, an ActiveRecord::ConnectionNotEstablished error will be raised.

When a shard and role is passed, Active Record will first lookup the role, and then look up the connection by shard key.

ActiveRecord::Base.connected_to(role: :reading, shard: :shard_one_replica) do
  Dog.first # finds first Dog record stored on the shard one replica
end


136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
# File 'activerecord/lib/active_record/connection_handling.rb', line 136

def connected_to(role: nil, shard: nil, prevent_writes: false, &blk)
  if self != Base && !abstract_class
    raise NotImplementedError, "calling `connected_to` is only allowed on ActiveRecord::Base or abstract classes."
  end

  if !connection_class? && !primary_class?
    raise NotImplementedError, "calling `connected_to` is only allowed on the abstract class that established the connection."
  end

  unless role || shard
    raise ArgumentError, "must provide a `shard` and/or `role`."
  end

  with_role_and_shard(role, shard, prevent_writes, &blk)
end

#connected_to?(role:, shard: ActiveRecord::Base.default_shard) ⇒ Boolean

Returns true if role is the current connected role and/or current connected shard. If no shard is passed, the default will be used.

ActiveRecord::Base.connected_to(role: :writing) do
  ActiveRecord::Base.connected_to?(role: :writing) #=> true
  ActiveRecord::Base.connected_to?(role: :reading) #=> false
end

ActiveRecord::Base.connected_to(role: :reading, shard: :shard_one) do
  ActiveRecord::Base.connected_to?(role: :reading, shard: :shard_one) #=> true
  ActiveRecord::Base.connected_to?(role: :reading, shard: :default) #=> false
  ActiveRecord::Base.connected_to?(role: :writing, shard: :shard_one) #=> true
end

Returns:

  • (Boolean)


253
254
255
# File 'activerecord/lib/active_record/connection_handling.rb', line 253

def connected_to?(role:, shard: ActiveRecord::Base.default_shard)
  current_role == role.to_sym && current_shard == shard.to_sym
end

#connected_to_all_shards(role: nil, prevent_writes: false, &blk) ⇒ Object

Passes the block to connected_to for every shard the model is configured to connect to (if any), and returns the results in an array.

Optionally, role and/or prevent_writes can be passed which will be forwarded to each connected_to call.



186
187
188
189
190
# File 'activerecord/lib/active_record/connection_handling.rb', line 186

def connected_to_all_shards(role: nil, prevent_writes: false, &blk)
  shard_keys.map do |shard|
    connected_to(shard: shard, role: role, prevent_writes: prevent_writes, &blk)
  end
end

#connected_to_many(*classes, role:, shard: nil, prevent_writes: false) ⇒ Object

Connects a role and/or shard to the provided connection names. Optionally prevent_writes can be passed to block writes on a connection. reading will automatically set prevent_writes to true.

connected_to_many is an alternative to deeply nested connected_to blocks.

Usage:

ActiveRecord::Base.connected_to_many(AnimalsRecord, MealsRecord, role: :reading) do
  Dog.first # Read from animals replica
  Dinner.first # Read from meals replica
  Person.first # Read from primary writer
end


165
166
167
168
169
170
171
172
173
174
175
176
177
178
# File 'activerecord/lib/active_record/connection_handling.rb', line 165

def connected_to_many(*classes, role:, shard: nil, prevent_writes: false)
  classes = classes.flatten

  if self != Base || classes.include?(Base)
    raise NotImplementedError, "connected_to_many can only be called on ActiveRecord::Base."
  end

  prevent_writes = true if role == ActiveRecord.reading_role

  append_to_connected_to_stack(role: role, shard: shard, prevent_writes: prevent_writes, klasses: classes)
  yield
ensure
  connected_to_stack.pop
end

#connecting_to(role: default_role, shard: default_shard, prevent_writes: false) ⇒ Object

Use a specified connection.

This method is useful for ensuring that a specific connection is being used. For example, when booting a console in readonly mode.

It is not recommended to use this method in a request since it does not yield to a block like connected_to.



199
200
201
202
203
# File 'activerecord/lib/active_record/connection_handling.rb', line 199

def connecting_to(role: default_role, shard: default_shard, prevent_writes: false)
  prevent_writes = true if role == ActiveRecord.reading_role

  append_to_connected_to_stack(role: role, shard: shard, prevent_writes: prevent_writes, klasses: [self])
end

#connectionObject

Soft deprecated. Use #with_connection or #lease_connection instead.



274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
# File 'activerecord/lib/active_record/connection_handling.rb', line 274

def connection
  pool = connection_pool
  if pool.permanent_lease?
    case ActiveRecord.permanent_connection_checkout
    when :deprecated
      ActiveRecord.deprecator.warn <<~MESSAGE
        Called deprecated `ActiveRecord::Base.connection` method.

        Either use `with_connection` or `lease_connection`.
      MESSAGE
    when :disallowed
      raise ActiveRecordError, <<~MESSAGE
        Called deprecated `ActiveRecord::Base.connection` method.

        Either use `with_connection` or `lease_connection`.
      MESSAGE
    end
    pool.lease_connection
  else
    pool.active_connection
  end
end

#connection_db_configObject

Returns the db_config object from the associated connection:

ActiveRecord::Base.connection_db_config
  #<ActiveRecord::DatabaseConfigurations::HashConfig:0x00007fd1acbded10 @env_name="development",
    @name="primary", @config={pool: 5, timeout: 5000, database: "storage/development.sqlite3", adapter: "sqlite3"}>

Use only for reading.



334
335
336
# File 'activerecord/lib/active_record/connection_handling.rb', line 334

def connection_db_config
  connection_pool.db_config
end

#connection_poolObject



342
343
344
# File 'activerecord/lib/active_record/connection_handling.rb', line 342

def connection_pool
  connection_handler.retrieve_connection_pool(connection_specification_name, role: current_role, shard: current_shard, strict: true)
end

#connects_to(database: {}, shards: {}) ⇒ Object

Connects a model to the databases specified. The database keyword takes a hash consisting of a role and a database_key.

This will look up the database config using the database_key and establish a connection to that config.

class AnimalsModel < ApplicationRecord
  self.abstract_class = true

  connects_to database: { writing: :primary, reading: :primary_replica }
end

connects_to also supports horizontal sharding. The horizontal sharding API supports read replicas as well. You can connect a model to a list of shards like this:

class AnimalsModel < ApplicationRecord
  self.abstract_class = true

  connects_to shards: {
    default: { writing: :primary, reading: :primary_replica },
    shard_two: { writing: :primary_shard_two, reading: :primary_shard_replica_two }
  }
end

Returns an array of database connections.

Raises:

  • (NotImplementedError)


81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
# File 'activerecord/lib/active_record/connection_handling.rb', line 81

def connects_to(database: {}, shards: {})
  raise NotImplementedError, "`connects_to` can only be called on ActiveRecord::Base or abstract classes" unless self == Base || abstract_class?

  if database.present? && shards.present?
    raise ArgumentError, "`connects_to` can only accept a `database` or `shards` argument, but not both arguments."
  end

  connections = []

  @shard_keys = shards.keys

  if shards.empty?
    shards[:default] = database
  end

  self.default_shard = shards.keys.first

  shards.each do |shard, database_keys|
    database_keys.each do |role, database_key|
      db_config = resolve_config_for_connection(database_key)

      self.connection_class = true
      connections << connection_handler.establish_connection(db_config, owner_name: self, role: role, shard: shard.to_sym)
    end
  end

  connections
end

#establish_connection(config_or_env = nil) ⇒ Object

Establishes the connection to the database. Accepts a hash as input where the :adapter key must be specified with the name of a database adapter (in lower-case) example for regular databases (MySQL, PostgreSQL, etc):

ActiveRecord::Base.establish_connection(
  adapter:  "mysql2",
  host:     "localhost",
  username: "myuser",
  password: "mypass",
  database: "somedatabase"
)

Example for SQLite database:

ActiveRecord::Base.establish_connection(
  adapter:  "sqlite3",
  database: "path/to/dbfile"
)

Also accepts keys as strings (for parsing from YAML for example):

ActiveRecord::Base.establish_connection(
  "adapter"  => "sqlite3",
  "database" => "path/to/dbfile"
)

Or a URL:

ActiveRecord::Base.establish_connection(
  "postgres://myuser:mypass@localhost/somedatabase"
)

In case ActiveRecord::Base.configurations is set (Rails automatically loads the contents of config/database.yml into it), a symbol can also be given as argument, representing a key in the configuration hash:

ActiveRecord::Base.establish_connection(:production)

The exceptions AdapterNotSpecified, AdapterNotFound, and ArgumentError may be returned on an error.



50
51
52
53
54
# File 'activerecord/lib/active_record/connection_handling.rb', line 50

def establish_connection(config_or_env = nil)
  config_or_env ||= DEFAULT_ENV.call.to_sym
  db_config = resolve_config_for_connection(config_or_env)
  connection_handler.establish_connection(db_config, owner_name: self, role: current_role, shard: current_shard)
end

#lease_connectionObject

Returns the connection currently associated with the class. This can also be used to “borrow” the connection to do database work unrelated to any of the specific Active Records. The connection will remain leased for the entire duration of the request or job, or until #release_connection is called.



269
270
271
# File 'activerecord/lib/active_record/connection_handling.rb', line 269

def lease_connection
  connection_pool.lease_connection
end

#primary_class?Boolean

:nodoc:

Returns:

  • (Boolean)


323
324
325
# File 'activerecord/lib/active_record/connection_handling.rb', line 323

def primary_class? # :nodoc:
  self == Base || application_record_class?
end

#prohibit_shard_swapping(enabled = true) ⇒ Object

Prohibit swapping shards while inside of the passed block.

In some cases you may want to be able to swap shards but not allow a nested call to connected_to or connected_to_many to swap again. This is useful in cases you’re using sharding to provide per-request database isolation.



211
212
213
214
215
216
217
# File 'activerecord/lib/active_record/connection_handling.rb', line 211

def prohibit_shard_swapping(enabled = true)
  prev_value = ActiveSupport::IsolatedExecutionState[:active_record_prohibit_shard_swapping]
  ActiveSupport::IsolatedExecutionState[:active_record_prohibit_shard_swapping] = enabled
  yield
ensure
  ActiveSupport::IsolatedExecutionState[:active_record_prohibit_shard_swapping] = prev_value
end

#release_connectionObject

Return the currently leased connection into the pool



298
299
300
# File 'activerecord/lib/active_record/connection_handling.rb', line 298

def release_connection
  connection_pool.release_connection
end

#remove_connectionObject



355
356
357
358
359
360
361
362
363
364
365
366
# File 'activerecord/lib/active_record/connection_handling.rb', line 355

def remove_connection
  name = @connection_specification_name if defined?(@connection_specification_name)

  # if removing a connection that has a pool, we reset the
  # connection_specification_name so it will use the parent
  # pool.
  if connection_handler.retrieve_connection_pool(name, role: current_role, shard: current_shard)
    self.connection_specification_name = nil
  end

  connection_handler.remove_connection_pool(name, role: current_role, shard: current_shard)
end

#retrieve_connectionObject



346
347
348
# File 'activerecord/lib/active_record/connection_handling.rb', line 346

def retrieve_connection
  connection_handler.retrieve_connection(connection_specification_name, role: current_role, shard: current_shard)
end

#schema_cacheObject

:nodoc:



368
369
370
# File 'activerecord/lib/active_record/connection_handling.rb', line 368

def schema_cache # :nodoc:
  connection_pool.schema_cache
end

#shard_keysObject



376
377
378
# File 'activerecord/lib/active_record/connection_handling.rb', line 376

def shard_keys
  connection_class_for_self.instance_variable_get(:@shard_keys) || []
end

#shard_swapping_prohibited?Boolean

Determine whether or not shard swapping is currently prohibited

Returns:

  • (Boolean)


220
221
222
# File 'activerecord/lib/active_record/connection_handling.rb', line 220

def shard_swapping_prohibited?
  ActiveSupport::IsolatedExecutionState[:active_record_prohibit_shard_swapping]
end

#sharded?Boolean

Returns:

  • (Boolean)


380
381
382
# File 'activerecord/lib/active_record/connection_handling.rb', line 380

def sharded?
  shard_keys.any?
end

#while_preventing_writes(enabled = true, &block) ⇒ Object

Prevent writing to the database regardless of role.

In some cases you may want to prevent writes to the database even if you are on a database that can write. while_preventing_writes will prevent writes to the database for the duration of the block.

This method does not provide the same protection as a readonly user and is meant to be a safeguard against accidental writes.

See READ_QUERY for the queries that are blocked by this method.



235
236
237
# File 'activerecord/lib/active_record/connection_handling.rb', line 235

def while_preventing_writes(enabled = true, &block)
  connected_to(role: current_role, prevent_writes: enabled, &block)
end

#with_connection(prevent_permanent_checkout: false, &block) ⇒ Object

Checkouts a connection from the pool, yield it and then check it back in. If a connection was already leased via #lease_connection or a parent call to #with_connection, that same connection is yieled. If #lease_connection is called inside the block, the connection won’t be checked back in. If #connection is called inside the block, the connection won’t be checked back in unless the prevent_permanent_checkout argument is set to true.



309
310
311
# File 'activerecord/lib/active_record/connection_handling.rb', line 309

def with_connection(prevent_permanent_checkout: false, &block)
  connection_pool.with_connection(prevent_permanent_checkout: prevent_permanent_checkout, &block)
end