Module: ActiveRecord::ConnectionHandling

Included in:
Base
Defined in:
lib/active_record/connection_adapters/mysql2_adapter.rb,
lib/active_record/connection_handling.rb,
lib/active_record/connection_adapters/sqlite3_adapter.rb,
lib/active_record/connection_adapters/postgresql_adapter.rb

Overview

:nodoc:

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_name ⇒ Object

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



286
287
288
289
290
291
 
# File 'lib/active_record/connection_handling.rb', line 286

def connection_specification_name
  if !defined?(@connection_specification_name) || @connection_specification_name.nil?
    return self == Base ? Base.name : superclass.connection_specification_name
  end
  @connection_specification_name
end

Instance Method Details

#clear_cache! ⇒ Object

:nodoc:



333
334
335
 
# File 'lib/active_record/connection_handling.rb', line 333

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

#clear_query_caches_for_current_thread ⇒ Object

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



266
267
268
269
270
271
272
273
274
 
# File 'lib/active_record/connection_handling.rb', line 266

def clear_query_caches_for_current_thread
  if ActiveRecord.legacy_connection_handling
    ActiveRecord::Base.connection_handlers.each_value do |handler|
      clear_on_handler(handler)
    end
  else
    clear_on_handler(ActiveRecord::Base.connection_handler)
  end
end

#connected? ⇒ Boolean

Returns true if Active Record is connected.

Returns:

  • (Boolean) 


317
318
319
 
# File 'lib/active_record/connection_handling.rb', line 317

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



137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
 
# File 'lib/active_record/connection_handling.rb', line 137

def connected_to(role: nil, shard: nil, prevent_writes: false, &blk)
  if ActiveRecord.legacy_connection_handling
    if self != Base
      raise NotImplementedError, "`connected_to` can only be called on ActiveRecord::Base with legacy connection handling."
    end
  else
    if self != Base && !abstract_class
      raise NotImplementedError, "calling `connected_to` is only allowed on ActiveRecord::Base or abstract classes."
    end

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

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

Returns:

  • (Boolean) 


252
253
254
 
# File 'lib/active_record/connection_handling.rb', line 252

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



172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
 
# File 'lib/active_record/connection_handling.rb', line 172

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

  if ActiveRecord.legacy_connection_handling
    raise NotImplementedError, "connected_to_many is not available with legacy connection handling"
  end

  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.



198
199
200
201
202
203
204
205
206
 
# File 'lib/active_record/connection_handling.rb', line 198

def connecting_to(role: default_role, shard: default_shard, prevent_writes: false)
  if ActiveRecord.legacy_connection_handling
    raise NotImplementedError, "`connecting_to` is not available with `legacy_connection_handling`."
  end

  prevent_writes = true if role == ActiveRecord.reading_role

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

#connection ⇒ Object

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.



279
280
281
 
# File 'lib/active_record/connection_handling.rb', line 279

def connection
  retrieve_connection
end

#connection_db_config ⇒ Object

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: "db/development.sqlite3", adapter: "sqlite3"}>

Use only for reading.



304
305
306
 
# File 'lib/active_record/connection_handling.rb', line 304

def connection_db_config
  connection_pool.db_config
end

#connection_pool ⇒ Object 



308
309
310
 
# File 'lib/active_record/connection_handling.rb', line 308

def connection_pool
  connection_handler.retrieve_connection_pool(connection_specification_name, role: current_role, shard: current_shard) || raise(ConnectionNotEstablished)
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 create a connection handler for switching between connections, look up the config hash using the database_key and finally establishes 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 also supports read replicas. 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
109
 
# File '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 = []

  database.each do |role, database_key|
    db_config, owner_name = resolve_config_for_connection(database_key)
    handler = lookup_connection_handler(role.to_sym)

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

  shards.each do |shard, database_keys|
    database_keys.each do |role, database_key|
      db_config, owner_name = resolve_config_for_connection(database_key)
      handler = lookup_connection_handler(role.to_sym)

      self.connection_class = true
      connections << handler.establish_connection(db_config, owner_name: owner_name, 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.



49
50
51
52
53
 
# File 'lib/active_record/connection_handling.rb', line 49

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

#lookup_connection_handler(handler_key) ⇒ Object

:nodoc:



256
257
258
259
260
261
262
263
 
# File 'lib/active_record/connection_handling.rb', line 256

def lookup_connection_handler(handler_key) # :nodoc:
  if ActiveRecord.legacy_connection_handling
    handler_key ||= ActiveRecord.writing_role
    connection_handlers[handler_key] ||= ActiveRecord::ConnectionAdapters::ConnectionHandler.new
  else
    ActiveRecord::Base.connection_handler
  end
end

#mysql2_connection(config) ⇒ Object

Establishes a connection to the database that’s used by all Active Record objects.



12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
 
# File 'lib/active_record/connection_adapters/mysql2_adapter.rb', line 12

def mysql2_connection(config)
  config = config.symbolize_keys
  config[:flags] ||= 0

  if config[:flags].kind_of? Array
    config[:flags].push "FOUND_ROWS"
  else
    config[:flags] |= Mysql2::Client::FOUND_ROWS
  end

  ConnectionAdapters::Mysql2Adapter.new(
    ConnectionAdapters::Mysql2Adapter.new_client(config),
    logger,
    nil,
    config,
  )
end

#postgresql_connection(config) ⇒ Object

Establishes a connection to the database that’s used by all Active Record objects



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
 
# File 'lib/active_record/connection_adapters/postgresql_adapter.rb', line 25

def postgresql_connection(config)
  conn_params = config.symbolize_keys.compact

  # Map ActiveRecords param names to PGs.
  conn_params[:user] = conn_params.delete(:username) if conn_params[:username]
  conn_params[:dbname] = conn_params.delete(:database) if conn_params[:database]

  # Forward only valid config params to PG::Connection.connect.
  valid_conn_param_keys = PG::Connection.conndefaults_hash.keys + [:requiressl]
  conn_params.slice!(*valid_conn_param_keys)

  ConnectionAdapters::PostgreSQLAdapter.new(
    ConnectionAdapters::PostgreSQLAdapter.new_client(conn_params),
    logger,
    conn_params,
    config,
  )
end

#primary_class? ⇒ Boolean

:nodoc:

Returns:

  • (Boolean) 


293
294
295
 
# File 'lib/active_record/connection_handling.rb', line 293

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.



214
215
216
217
218
219
220
 
# File 'lib/active_record/connection_handling.rb', line 214

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

#remove_connection(name = nil) ⇒ Object 



321
322
323
324
325
326
327
328
329
330
331
 
# File 'lib/active_record/connection_handling.rb', line 321

def remove_connection(name = nil)
  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_connection ⇒ Object 



312
313
314
 
# File 'lib/active_record/connection_handling.rb', line 312

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

#shard_swapping_prohibited? ⇒ Boolean

Determine whether or not shard swapping is currently prohibited

Returns:

  • (Boolean) 


223
224
225
 
# File 'lib/active_record/connection_handling.rb', line 223

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

#sqlite3_connection(config) ⇒ Object 



18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
 
# File 'lib/active_record/connection_adapters/sqlite3_adapter.rb', line 18

def sqlite3_connection(config)
  config = config.symbolize_keys

  # Require database.
  unless config[:database]
    raise ArgumentError, "No database file specified. Missing argument: database"
  end

  # Allow database path relative to Rails.root, but only if the database
  # path is not the special path that tells sqlite to build a database only
  # in memory.
  if ":memory:" != config[:database] && !config[:database].to_s.start_with?("file:")
    config[:database] = File.expand_path(config[:database], Rails.root) if defined?(Rails.root)
    dirname = File.dirname(config[:database])
    Dir.mkdir(dirname) unless File.directory?(dirname)
  end

  db = SQLite3::Database.new(
    config[:database].to_s,
    config.merge(results_as_hash: true)
  )

  ConnectionAdapters::SQLite3Adapter.new(db, logger, nil, config)
rescue Errno::ENOENT => error
  if error.message.include?("No such file or directory")
    raise ActiveRecord::NoDatabaseError
  else
    raise
  end
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.



238
239
240
241
242
243
244
 
# File 'lib/active_record/connection_handling.rb', line 238

def while_preventing_writes(enabled = true, &block)
  if ActiveRecord.legacy_connection_handling
    connection_handler.while_preventing_writes(enabled, &block)
  else
    connected_to(role: current_role, prevent_writes: enabled, &block)
  end
end