Navigation

Configuration

Mongoid is customarily configured through a mongoid.yml file that specifies options and clients. The simplest configuration is as follows, which configures Mongoid to talk to a MongoDB server at “localhost:27017” and use the database named “mongoid”.

development:
  clients:
    default:
      database: mongoid
      hosts:
        - localhost:27017

The top level key in the configuration file, development in the above example, refers to the environment name which the application is executing in, i.e. development, test or production. The third level key, default in the above example, refers to the Mongo client name. Most applications will use a single client named default.

Generating Default Configuration

If you are using Ruby on Rails, you can have Mongoid generate a default configuration file for you by running the following command:

rails g mongoid:config

The configuration file will be placed in config/mongoid.yml. An initializer will also be created and placed in config/initializers/mongoid.rb. It is recommended that all configuration be specified in config/mongoid.yml, but if you prefer, the mongoid.rb initializer may also be used to set configuration options. Note, though, that settings in mongoid.yml always take precedence over settings in the initializer.

If you are not using Ruby on Rails, you can copy the minimal configuration given above and save it as config/mongoid.yml.

Loading Mongoid Configuration

If you are using Ruby on Rails, Mongoid configuration is automatically loaded for the current environment as stored in Rails.env when the application loads.

You may need to configure the ORM for your application to be Mongoid by adding the following to application.rb:

config.generators do |g|
  g.orm :mongoid
end

If you are not using Ruby on Rails, Mongoid configuration must be loaded manually. This can be done via the Mongoid.load! method, which takes the configuration file path as its argument, as follows:

# Use automatically detected environment name
Mongoid.load!("path/to/your/mongoid.yml")

# Specify environment name manually
Mongoid.load!("path/to/your/mongoid.yml", :production)

When Mongoid is asked to automatically detect the environment name, it does so by examining the following sources, in order:

  • If Rails top level constant is defined, Rails.env.
  • If Sinatra top level constant is defined, Sinatra::Base.environment.
  • The RACK_ENV environment variable.
  • The MONGOID_ENV environment variable.

It is also possible to configure Mongoid directly in Ruby, without using a configuration file. This configuration style does not support the concept of environments - whatever configuration is provided, it is applied to the current environment - but it does support defining multiple clients.

Mongoid.configure do |config|
  config.clients.default = {
    hosts: ['localhost:27017'],
    database: 'my_db',
  }

  config.log_level = :warn
end

Note

Mongoid must be configured before any component of it is used or referenced. Once a component is used or referenced, changing configuration may not apply changes to already instantiated components.

Mongoid Configuration Options

The following annotated example mongoid.yml demonstrates how Mongoid can be configured.

Mongoid delegates to the Ruby driver for client configuration. Please review the driver documentation for details on driver options.

development:
  # Configure available database clients. (required)
  clients:
    # Define the default client. (required)
    default:
      # A uri may be defined for a client:
      # uri: 'mongodb://user:password@myhost1.mydomain.com:27017/my_db'
      # Please see driver documentation for details. Alternatively, you can
      # define the following:
      #
      # Define the name of the default database that Mongoid can connect to.
      # (required).
      database: my_db
      # Provide the hosts the default client can connect to. Must be an array
      # of host:port pairs. (required)
      hosts:
        - myhost1.mydomain.com:27017
        - myhost2.mydomain.com:27017
        - myhost3.mydomain.com:27017
      options:
        # These options are Ruby driver options, documented in
        # https://mongodb.com/docs/ruby-driver/current/reference/create-client/

        # Change the default write concern. (default = { w: 1 })
        write:
          w: 1

        # Change the default read preference. Valid options for mode are: :secondary,
        # :secondary_preferred, :primary, :primary_preferred, :nearest
        # (default: primary)
        read:
          mode: :secondary_preferred
          tag_sets:
            - use: web

        # The name of the user for authentication.
        user: 'user'

        # The password of the user for authentication.
        password: 'password'

        # The user's database roles.
        roles:
          - 'dbOwner'

        # Change the default authentication mechanism. Please see the
        # driver documentation linked above for details on how to configure
        # authentication. Valid options are :aws, :gssapi, :mongodb_cr,
        # :mongodb_x509, :plain, :scram and :scram256 (default on 3.0
        # and higher servers is :scram, default on 2.6 servers is :plain)
        auth_mech: :scram

        # Specify the auth source, i.e. the database or other source which
        # contains the user's login credentials. Allowed values for auth source
        # depend on the authentication mechanism, as explained in the server documentation:
        # https://mongodb.com/docs/manual/reference/connection-string/#mongodb-urioption-urioption.authSource
        # If no auth source is specified, the default auth source as
        # determined by the driver will be used. Please refer to:
        # https://mongodb.com/docs/ruby-driver/current/reference/authentication/#auth-source
        auth_source: admin

        # Connect directly to and perform all operations on the specified
        # server, bypassing replica set node discovery and monitoring.
        # Exactly one host address must be specified. (default: false)
        #direct_connection: true

        # Deprecated. Force the driver to connect in a specific way instead
        # of automatically discovering the deployment type and connecting
        # accordingly. To connect directly to a replica set node bypassing
        # node discovery and monitoring, use direct_connection: true instead
        # of this option. Possible values: :direct, :replica_set, :sharded.
        # (default: none)
        #connect: :direct

        # Change the default time in seconds the server monitors refresh their status
        # via hello commands. (default: 10)
        heartbeat_frequency: 10

        # The time in seconds for selecting servers for a near read preference. (default: 0.015)
        local_threshold: 0.015

        # The timeout in seconds for selecting a server for an operation. (default: 30)
        server_selection_timeout: 30

        # The maximum number of connections in the connection pool. (default: 5)
        max_pool_size: 5

        # The minimum number of connections in the connection pool. (default: 1)
        min_pool_size: 1

        # The time to wait, in seconds, in the connection pool for a connection
        # to be checked in before timing out. (default: 1)
        wait_queue_timeout: 1

        # The time to wait to establish a connection before timing out, in seconds.
        # (default: 10)
        connect_timeout: 10

        # How long to wait for a response for each operation sent to the
        # server. This timeout should be set to a value larger than the
        # processing time for the longest operation that will be executed
        # by the application. Note that this is a client-side timeout;
        # the server may continue executing an operation after the client
        # aborts it with the SocketTimeout exception.
        # (default: nil, meaning no timeout)
        socket_timeout: 5

        # The name of the replica set to connect to. Servers provided as seeds that do
        # not belong to this replica set will be ignored.
        replica_set: my_replica_set

        # Whether to connect to the servers via ssl. (default: false)
        ssl: true

        # The certificate file used to identify the connection against MongoDB.
        ssl_cert: /path/to/my.cert

        # The private keyfile used to identify the connection against MongoDB.
        # Note that even if the key is stored in the same file as the certificate,
        # both need to be explicitly specified.
        ssl_key: /path/to/my.key

        # A passphrase for the private key.
        ssl_key_pass_phrase: password

        # Whether or not to do peer certification validation. (default: true)
        ssl_verify: true

        # The file containing a set of concatenated certification authority certifications
        # used to validate certs passed from the other end of the connection.
        ssl_ca_cert: /path/to/ca.cert

        # Compressors to use. (default is to not use compression)
        compressors: [zlib]

  # Configure Mongoid-specific options. (optional)
  options:
    # Application name that is printed to the MongoDB logs upon establishing
    # a connection in server versions 3.4 or greater. Note that the name
    # cannot exceed 128 bytes in length. It is also used as the database name
    # if the database name is not explicitly defined. (default: nil)
    app_name: MyApplicationName

    # Type of executor for queries scheduled using ``load_async`` method.
    #
    # There are two possible values for this option:
    #
    #   - :immediate - Queries will be immediately executed on a current thread.
    #       This is the default option.
    #   - :global_thread_pool - Queries will be executed asynchronously in
    #       background using a thread pool.
    #async_query_executor: :immediate

    # Mark belongs_to associations as required by default, so that saving a
    # model with a missing belongs_to association will trigger a validation
    # error. (default: true)
    belongs_to_required_by_default: true

    # Set the global discriminator key. (default: "_type")
    discriminator_key: "_type"

    # Raise an exception when a field is redefined. (default: false)
    duplicate_fields_exception: false

    # Defines how many asynchronous queries can be executed concurrently.
    # This option should be set only if `async_query_executor` option is set
    # to `:global_thread_pool`.
    #global_executor_concurrency: nil

    # Include the root model name in json serialization. (default: false)
    include_root_in_json: false

    # Include the _type field in serialization. (default: false)
    include_type_for_serialization: false

    # Whether to join nested persistence contexts for atomic operations
    # to parent contexts by default. (default: false)
    join_contexts: false

    # When this flag is false, a document will become read-only only once the
    # #readonly! method is called, and an error will be raised on attempting
    # to save or update such documents, instead of just on delete. When this
    # flag is true, a document is only read-only if it has been projected
    # using #only or #without, and read-only documents will not be
    # deletable/destroyable, but will be savable/updatable.
    # When this feature flag is turned on, the read-only state will be reset on
    # reload, but when it is turned off, it won't be.
    # (default: false)
    #legacy_readonly: true

    # Set the Mongoid and Ruby driver log levels when Mongoid is not using
    # Ruby on Rails logger instance. (default: :info)
    log_level: :info

    # When using the BigDecimal field type, store the value in the database
    # as a BSON::Decimal128 instead of a string. (default: true)
    #map_big_decimal_to_decimal128: true

    # Preload all models in development, needed when models use
    # inheritance. (default: false)
    preload_models: false

    # Raise an error when performing a #find and the document is not found.
    # (default: true)
    raise_not_found_error: true

    # Raise an error when defining a scope with the same name as an
    # existing method. (default: false)
    scope_overwrite_exception: false

    # Return stored times as UTC. See the time zone section below for
    # further information. Most applications should not use this option.
    # (default: false)
    use_utc: false

    # (Deprecated) In MongoDB 4.0 and earlier, set whether to create
    # indexes in the background by default. (default: false)
    background_indexing: false

  # Configure driver-specific options. (optional)
  driver_options:
    # When this flag is turned off, inline options will be correctly
    # propagated to Mongoid and Driver finder methods. When this flag is turned
    # on those options will be ignored. For example, with this flag turned
    # off, Band.all.limit(1).count will take the limit into account, while
    # when this flag is turned on, that limit is ignored. The affected driver
    # methods are: aggregate, count, count_documents, distinct, and
    # estimated_document_count. The corresponding Mongoid methods are also
    # affected. (default: false, driver version: 2.18.0+)
    #broken_view_options: false

    # Validates that there are no atomic operators (those that start with $)
    # in the root of a replacement document, and that there are only atomic
    # operators at the root of an update document. If this feature flag is on,
    # an error will be raised on an invalid update or replacement document,
    # if not, a warning will be output to the logs. This flag does not affect
    # Mongoid as of 8.0, but will affect calls to driver update/replace
    # methods. (default: false, driver version: 2.18.0+)
    #validate_update_replace: false

Version Based Defaults

Mongoid supports setting the configuration options to the defaults for specific versions. This is useful for upgrading to a new Mongoid version. When upgrading your Mongoid version, the following should be set on Mongoid::Config:

Mongoid.configure do |config|
  config.load_defaults <OLD VERSION>
end

This way, when upgrading to a new version of Mongoid, your code will run with the configuration options from the previous version of Mongoid. Then, one-by-one, you can change the feature flags for the new version, and test that your code still acts as expected. Once all of the new feature flags have been accounted for, the call to load_defaults may be changed to take in the new version, and all of the changed feature flags may be removed.

For example, suppose we’re upgrading from 7.5 to 8.0. Between these two versions, two feature flags were added: legacy_attributes and map_big_decimal_to_decimal128. Before upgrading to Mongoid 8, add the following to your Mongoid::Config:

Mongoid.configure do |config|
  config.load_defaults 7.5
end

After upgrading to Mongoid 8.0 in your Gemfile, any feature flags will remain set to their 7.5 default behavior: legacy_attributes: true, map_big_decimal_to_decimal128: false. You may then flip these feature flags one-by-one to their 8.0 behavior:

Mongoid.configure do |config|
  config.load_defaults 7.5
  config.legacy_attributes = false
  # config.map_big_decimal_to_decimal128 = true
end

We recommend do these one at a time, so in the example above we leave the second flag commented out. After verifying your code works as expected with the legacy_attributes flag turned off, the map_big_decimal_to_decimal128 setting can be uncommented. Once that functionality is verified as well, both of those lines can be removed and the load_defaults replaced with:

Mongoid.configure do |config|
  config.load_defaults 8.0
end

ERb Preprocessing

When loading a configuration file, Mongoid processes it with ERb before parsing it as YAML. This allows, for example, constructing the contents of the configuration file at runtime based on environment variables:

development:
  clients:
    default:
      uri: "<%= ENV['MONGODB_URI'] %>"

Note

When outputting values from ERb, ensure the values are valid YAML and escape them as needed.

Note

Since ERb rendering is performed prior to YAML parsing, all ERb directives in the configuration file are evaluated, including those occurring in YAML comments.

Logging

When configuring logging, it is important to keep in mind that Mongoid provides a model layer on top of the MongoDB Ruby driver, and the driver dispatches the CRUD operations to the MongoDB deployment. Therefore, some of the logging output in an application using Mongoid comes from Mongoid itself, and some comes from the driver.

The Mongo client is a Ruby driver client instance, therefore the logger of a Mongo client is the Ruby driver logger, not the Mongoid logger. In other words:

# Ruby driver logger, not Mongoid logger
Mongoid.client(:default).logger

Depending on whether Mongoid is used in a Ruby on Rails application, and how both Mongoid and Ruby driver are configured, they may use the same logger instance or different instances, potentially with different configurations.

In Ruby on Rails Application

When used in a Ruby on Rails application, Mongoid by default inherits the logger and the log level from Rails, and sets the driver’s logger to the same logger instance:

Rails.logger === Mongoid.logger
# => true

Mongoid.logger === Mongo::Logger.logger
# => true

To change the log level, use standard Rails configuration. Place the following in one of environment configuration files, such as config/environments/production.rb:

Rails.application.configure do
  config.log_level = :debug
end

Note

The log_level Mongoid configuration option is not used when Mongoid operates in a Rails application, because Mongoid inherits Rails’ log level in this case.

To configure either Mongoid or driver logger differently from the Rails logger, use an initializer as follows:

Rails.application.configure do
  config.after_initialize do
    # Change Mongoid log destination and/or level
    Mongoid.logger = Logger.new(STDERR).tap do |logger|
      logger.level = Logger::DEBUG
    end

    # Change driver log destination and/or level
    Mongo::Logger.logger = Logger.new(STDERR).tap do |logger|
      logger.level = Logger::DEBUG
    end
  end
end

Note

There is currently no provision in the Ruby standard library Logger to return the log device (i.e. the IO object) that a logger is using. To have, for example, Mongoid and/or the Ruby driver log to the standard Rails log file (e.g. log/development.log) but with a different level from standard Rails logger (Rails.logger), the file must be opened separately and the resulting IO object passed to the Logger constructor.

Note

Since by default Mongoid sets its own logger and the driver’s logger to the same instance as the Rails logger, modifying any of the instances affects all of them. For example the following changes log level for all three loggers, unless the application assigned a separate Logger instance to Mongo::Logger.logger as described above:

Mongoid::Logger.logger.level = Logger::DEBUG

Standalone

When not loaded in a Ruby on Rails application, Mongoid respects the log_level top level configuration option. It can be given in the configuration file as follows:

development:
  clients:
    default:
      # ...
  options:
    log_level: :debug

… or when configuring Mongoid inline:

Mongoid.configure do |config|
  config.log_level = :debug
end

The default log destination in Mongoid 7.1 and higher is standard error. The default log destination in Mongoid 7.0 and lower is standard output. To change the log destination, create a new logger instance as follows:

Mongoid.logger = Logger.new(STDERR).tap do |logger|
  logger.level = Logger::DEBUG
end

To change the Ruby driver log level or destination:

Mongo::Logger.logger = Logger.new(STDERR).tap do |logger|
  logger.level = Logger::DEBUG
end

To set the driver logger to be the same as the Mongoid logger:

Mongo::Logger.logger = Mongoid.logger

Note

Mongoid does not alter the driver’s logger when running in standalone mode.

Time Zones

Mongoid uses ActiveSupport’s time zone functionality, which is far more robust than Ruby’s standard library. Importantly, ActiveSupport allows configuration of Time.zone, a thread-global variable which provides context for working with date and time values.

While a thorough treatment of time zones in Ruby is outside the scope of this tutorial, the easiest and most reliable way of achieving correct time zone handling is as follows:

  1. Set the operating system’s time zone to UTC. For example, on Linux:
cp /usr/share/zoneinfo/UTC /etc/localtime
  1. Set your application default time zone to UTC:
# If using Rails, in application.rb:
class Application < Rails::Application
  config.time_zone = 'UTC'
end

# If not using Rails:
Time.zone = 'UTC'
  1. In each controller and job class, set the appropriate time zone in a before_filter at the earliest possible stage. As an example, if each user of your application can set their own time zone, you may wish to do:
class ApplicationController < ActionController::Base
  before_filter :fetch_user,
                :set_time_zone

  def set_time_zone
    Time.zone = @user.time_zone
  end
end
  1. From here, you may work with times naturally in the local time zone. For example, in a view:
Turned into a pumpkin after <%= cinderella.updated_at.seconds_after_midnight %> seconds!
  1. Use ActiveSupport methods instead of the Ruby standard library.

    • Time.zone.now or Time.current` instead of ``Time.now
    • Date.current instead of Date.today

    Critically, note that the latter Ruby standard library methods reference your system time zone (e.g. UTC) and not the value of Time.zone. As it is very easy to mistake these similarly named methods, we recommend to use Rubocop’s Rails/TimeZone cop in your CI.

Setting time zone on data loaded from MongoDB

MongoDB stores all times in UTC without time zone information. Mongoid models load and returns time values as instances of ActiveSupport::TimeWithZone. You may set the use_utc option to control how Mongoid sets the time zone when loading from the database:

  • If false (default), Mongoid will use Time.zone to set the time zone of time values are loaded from database.
  • If true, Mongoid will always set the time zone as UTC on loaded time values.

use_utc only affects how data is loaded, and does not affect how data is persisted. For example, if you assign a Time or ActiveSupport::TimeWithZone instance to a time field, the time zone information of the assigned instance always will be used irrespective of the use_utc setting. Alternatively, if you assign a string value to a time field, any time zone information in the string will be used if present. If the string does not include time zone information it will be parsed according to Time.zone. To illustrate:

Time.use_zone("Asia/Kolkata") do

  # String does not include time zone, so "Asia/Kolkata" will be used
  ghandi.born_at = "1869-10-02 7:10 PM"

  # Time zone in string (-0600) will be used
  amelia.born_at = "1897-07-24 11:30 -0600"
end

Configuring SSLContext

It may be desirable to further configure TLS options in your application, for example by enabling or disabling certain ciphers.

This can be done by setting TLS context hooks on the Ruby driver – TLS context hooks are user-provided Proc``s that will be invoked before any TLS socket connection in the driver and can be used to modify the underlying ``OpenSSL::SSL::SSLContext object used by the socket.

To set TLS context hooks, add Proc``s to the ``Mongo.tls_context_hooks array. This can be done in an initializer. The example below adds a hook that only enables the “AES256-SHA” cipher.

Mongo.tls_context_hooks.push(
  Proc.new { |context|
    context.ciphers = ["AES256-SHA"]
  }
)

# Only the AES256-SHA cipher will be enabled from this point forward

Every Proc in Mongo.tls_context_hooks will be passed an OpenSSL::SSL::SSLContext object as its sole argument. These procs will be executed sequentially during socket creation.

Warning

TLS context hooks are global and will affect all Mongo::Client instances in an application.

For more information about TLS context hooks, including best practices for assigning and removing them, see the Ruby driver documentation.

Client-Side Encryption

When loading the configuration file, Mongoid permits the file to contain BSON::Binary instances which are used for specifying keyId in the schema map for client-side encryption, as the following example shows:

development:
  clients:
    default:
      database: blog_development
      hosts: [localhost:27017]
      options:
        auto_encryption_options:
          key_vault_namespace: 'keyvault.datakeys'
          kms_providers:
            local:
              key: "z7iYiYKLuYymEWtk4kfny1ESBwwFdA58qMqff96A8ghiOcIK75lJGPUIocku8LOFjQuEgeIP4xlln3s7r93FV9J5sAE7zg8U"
          schema_map:
            blog_development.comments:
              properties:
                message:
                  encrypt:
                    keyId:
                      - !ruby/object:BSON::Binary
                        data: !binary |-
                          R/AgNcxASFiiJWKXqWGo5w==
                        type: :uuid
                    bsonType: "string"
                    algorithm: "AEAD_AES_256_CBC_HMAC_SHA_512-Deterministic"
              bsonType: "object"

Usage with Forking Servers

When using Mongoid with a forking web server such as Puma, Unicorn or Passenger, it is recommended to not perform any operations on Mongoid models in the parent process prior to the fork.

When a process forks, Ruby threads are not transferred to the child processes and the Ruby driver Client objects lose their background monitoring. The application will typically seem to work just fine until the deployment state changes (for example due to network errors, a maintenance event) at which point the application is likely to start getting NoServerAvailable exception when performing MongoDB operations.

If the parent process needs to perform operations on the MongoDB database, reset all clients in the workers after they forked. How to do so depends on the web server being used.

If the parent process does not need to perform operations on the MongoDB database after child processes are forked, close the clients in the parent prior to forking children. If the parent process performs operations on a Mongo client and does not close it, the parent process will continue consuming a connection slot in the cluster and will continue monitoring the cluster for as long as the parent remains alive.

Note

The close/reconnect pattern described here should be used with Ruby driver version 2.6.2 or higher. Previous driver versions did not recreate monitoring threads when reconnecting.

Puma

Use the on_worker_boot hook to reconnect clients in the workers and the before_fork hook to close clients in the parent process (Puma documentation):

on_worker_boot do
  if defined?(Mongoid)
    Mongoid::Clients.clients.each do |name, client|
      client.close
      client.reconnect
    end
  else
    raise "Mongoid is not loaded. You may have forgotten to enable app preloading."
  end
end

before_fork do
  if defined?(Mongoid)
    Mongoid.disconnect_clients
  end
end

Unicorn

Use the after_fork hook to reconnect clients in the workers and the before_fork hook to close clients in the parent process (Unicorn documentation):

after_fork do |server, worker|
  if defined?(Mongoid)
    Mongoid::Clients.clients.each do |name, client|
      client.close
      client.reconnect
    end
  else
    raise "Mongoid is not loaded. You may have forgotten to enable app preloading."
  end
end

before_fork do |server, worker|
  if defined?(Mongoid)
    Mongoid.disconnect_clients
  end
end

Passenger

Use the starting_worker_process hook to reconnect clients in the workers (Passenger documentation). Passenger does not appear to have a hook that is invoked in the parent process before the workers are forked.

if defined?(PhusionPassenger)
  PhusionPassenger.on_event(:starting_worker_process) do |forked|
    if forked
      Mongoid::Clients.clients.each do |name, client|
        client.close
        client.reconnect
      end
    end
  end
end

Query Cache Middleware

Enabling Query Cache for Rack Web Requests

The MongoDB Ruby Driver provides a Rack middleware which enables the Query Cache for the duration of each web request. Below is an example of how to enable the Query Cache Middleware in a Ruby on Rails application:

# config/application.rb

# Add Mongo::QueryCache::Middleware at the bottom of the middleware stack
# or before other middleware that queries MongoDB.
config.middleware.use Mongo::QueryCache::Middleware

Please refer to the Rails on Rack guide for more information about using Rack middleware in Rails applications.

Enabling Query Cache for ActiveJob

The MongoDB Ruby Driver also provides Query Cache middleware for ActiveJob. You may enable it for all jobs in an initializer:

# config/initializers/active_job.rb

# Enable Mongo driver query cache for ActiveJob
ActiveSupport.on_load(:active_job) do
  include Mongo::QueryCache::Middleware::ActiveJob
end

Or for a specific job class:

class MyJob < ActiveJob::Base
  include Mongo::QueryCache::Middleware::ActiveJob
end

Development Configuration

Driver’s default configuration is suitable for production deployment. In development, some settings can be adjusted to provide a better developer experience.

  • :server_selection_timeout: set this to a low value (e.g., 1) if your MongoDB server is running locally and you start it manually. A low server selection timeout will cause the driver to fail quickly when there is no server running.

Sample recommended development configuration:

development:
  clients:
    default:
      database: mongoid
      hosts:
        - localhost:27017
      options:
        server_selection_timeout: 1