Creating a Client

Using ::Mongo::Client

To connect to a MongoDB deployment, create a ::Mongo::Client object. Provide a list of hosts and options or a connection string URI </reference/connection-string/> to the::Mongo::Client constructor. The client's selected database defaults to admin.

By default, the driver will automatically detect the topology used by the deployment and connect appropriately.

To connect to a local standalone MongoDB deployment, specify the host and port of the server. In most cases you would also specify the database name to connect to; if no database name is specified, the client will use the admin database:

Mongo::Client.new([ '127.0.0.1:27017' ], database: 'mydb')

# Or using the URI syntax:
Mongo::Client.new("mongodb://127.0.0.1:27017/mydb")

To connect to MongoDB Atlas, specify the Atlas deployment URI:

Mongo::Client.new("mongodb+srv://username:myRealPassword@cluster0.mongodb.net/mydb?w=majority")

The driver will discover all nodes in the cluster and connect to them as needed.

Block Syntax

Another way to create a Mongo::Client object is to use the block syntax:

Mongo::Client.new(...) do |client|
  # work with the client
end

Note that when creating a client using this syntax, the client is automatically closed after the block finishes executing.

Database Selection

By default, the client will connect to the admin database.

The admin database is a special database in MongoDB often used for administrative tasks and storing administrative data such as users and roles (although users and roles may also be defined in other databases). In a sharded cluster, the admin database exists on the config servers </core/sharded-cluster-config-servers/#read-and-write-operations-on-config-servers> rather than the shard servers. Although it is possible to use the admin database for ordinary operations (such as storing application data), this is not recommended and the application should explicitly specify the database it wishes to use.

The database can be specified during Client construction:

# Using Ruby client options:
client = Mongo::Client.new(['localhost'], database: 'mydb')

# Using a MongoDB URI:
client = Mongo::Client.new('mongodb://localhost/mydb')

Given a Client instance, the use method can be invoked to obtain a new Client instance configured with the specified database:

client = Mongo::Client.new(['localhost'], database: 'mydb')

admin_client = client.use('admin')

# Issue an administrative command
admin_client.database.command(replSetGetConfig: 1).documents.first

There are other special databases in MongoDB which should be only used for their stated purposes:

• The config </reference/config-database/> database.
• The local </reference/local-database/> database.
• The $external database, which is used with PLAIN <plain>, Kerberos <kerberos> and X.509 <x.509> authentication mechanisms.

Connection Types

The driver will, by default, discover the type of deployment it is instructed to connect to (except for load-balanced deployments) and behave in the manner that matches the deployment type. The subsections below describe how the driver behaves in each of the deployment types as well as how to force particular behavior, bypassing automatic deployment type detection.

Note that the detection of deployment type happens when the driver receives the first reply from any of the servers it is instructed to connect to (unless the load-balancing mode is requested, see below). The driver will remain in the discovered or configured topology even if the underlying deployment is replaced by one of a different type. In particular, when replacing a replica set with a sharded cluster at the same address the client instance must be recreated (such as by restarting the application) for it to communicate with the sharded cluster.

Automatic discovery of load-balanced deployments is currently not supported. Load-balanced deployments will be treated as deployments of their underlying type, which would generally be sharded clusters. The driver will fail to correctly operate when treating a load-balanced deployment as a sharded cluster, therefore when the deployment is a load-balanced one the client must be explicitly configured to connect to a load balancer <load-balancer-connection>.

Standalone Server Connection

If the deployment is a single server, also known as a standalone deployment, all operations will be directed to the specified server.

If the server is shut down and replaced by a replica set node, the driver will continue sending all operations to that node, even if the node is or becomes a secondary.

To force a standalone connection, see the direct connection <direct-connection> section below.

Replica Set Connection [connect-replica-set]

When connecting to a replica set</replication/>, it is sufficient to pass the address of any node in the replica set to the driver. The node does not have to be the primary and it may be a hidden node. The driver will then automatically discover the remaining nodes.

However, it is recommended to specify all nodes that are part of the replica set, so that in the event of one or more nodes being unavailable (for example, due to maintenance or reconfiguration) the driver can still connect to the replica set.

Replica set connection examples:

Mongo::Client.new([ '127.0.0.1:27017' ], database: 'mydb')

Mongo::Client.new([ '127.0.0.1:27017', '127.0.0.1:27018' ], database: 'mydb')

# Or using the URI syntax:
Mongo::Client.new("mongodb://127.0.0.1:27017,127.0.0.1:27018/mydb")

To make the driver verify the replica set name upon connection, pass it using the replica_set Ruby option or the replicaSet URI option:

Mongo::Client.new([ '127.0.0.1:27017', '127.0.0.1:27018' ],
  database: 'mydb', replica_set: 'myapp')

# Or using the URI syntax:
Mongo::Client.new("mongodb://127.0.0.1:27017,127.0.0.1:27018/mydb?replicaSet=myapp")

If the deployment is not a replica set or uses a different replica set name, all operations will fail (until the expected replica set is returned by the servers).

It is also possible to force a replica set connection without specifying the replica set name. Doing so is generally unnecessary and is deprecated:

Mongo::Client.new([ '127.0.0.1:27017', '127.0.0.1:27018' ],
  database: 'mydb', connect: :replica_set)

# Or using the URI syntax:
Mongo::Client.new("mongodb://127.0.0.1:27017,127.0.0.1:27018/mydb?connect=replica_set")

To connect to a MongoDB Atlas cluster which is deployed as a replica set, connect to the URI:

Mongo::Client.new("mongodb+srv://username:myRealPassword@cluster0.mongodb.net/test?w=majority")

Please review the SRV URI notes <srv-uri-notes> if using SRV URIs.

Sharded Cluster Connection [connect-sharded-cluster]

To connect to a sharded cluster</sharding/> deployment, specify the addresses of the mongos routers:

Mongo::Client.new([ '1.2.3.4:27017', '1.2.3.5:27017' ], database: 'mydb')

Mongo::Client.new("mongodb://1.2.3.4:27017,1.2.3.5:27017/mydb")

Note that unlike a replica set connection, you may choose to connect to a subset of the mongos routers that exist in the deployment. The driver will monitor each router and will use the ones that are available (i.e., the driver will generally handle individual routers becoming unavailable due to failures or maintenance). When specifying the list of routers explicitly, the driver will not discover remaining routers that may be configured and will not attempt to connect to them.

The driver will automatically balance the operation load among the routers it is aware of.

To connect to a MongoDB Atlas cluster which is deployed as a sharded cluster, connect to the URI:

Mongo::Client.new("mongodb+srv://username:myRealPassword@cluster0.mongodb.net/test?w=majority")

When the driver connects to a sharded cluster via an SRV URI, it will periodically poll the SRV records of the address specified in the URI for changes and will automatically add and remove the mongos hosts to/from its list of servers as they are added and removed to/from the sharded cluster.

To force a sharded cluster connection, use the connect: :sharded option. Doing so is generally unnecessary and is deprecated:

Mongo::Client.new([ '127.0.0.1:27017', '127.0.0.1:27018' ],
  database: 'mydb', connect: :sharded)

# Or using the URI syntax:
Mongo::Client.new("mongodb://127.0.0.1:27017,127.0.0.1:27018/mydb?connect=sharded")

Please review the SRV URI notes <srv-uri-notes> if using SRV URIs.

Direct Connection

To disable the deployment type discovery and force all operations to be performed on a particular server, specify the direct_connection option:

Mongo::Client.new([ '1.2.3.4:27017' ], database: 'mydb', direct_connection: true)

# Or using the URI syntax:
Mongo::Client.new("mongodb://1.2.3.4:27017/mydb?directConnection=true")

Alternatively, the deprecated connect: :direct option is equivalent:

Mongo::Client.new([ '1.2.3.4:27017' ], database: 'mydb', connect: :direct)

# Or using the URI syntax:
Mongo::Client.new("mongodb://1.2.3.4:27017/mydb?connect=direct")

The direct connection mode is most useful for performing operations on a particular replica set node, although it also permits the underlying server to change type (e.g. from a replica set node to a mongos router, or vice versa).

Load Balancer Connection

Unlike other deployment types, the driver does not currently automatically detect a load-balanced deployment.

To connect to a load balancer, specify the load_balanced: true Ruby option or the loadBalanced=true URI option:

Mongo::Client.new([ '1.2.3.4:27017' ], database: 'mydb', load_balanced: true)

# Or using the URI syntax:
Mongo::Client.new("mongodb://1.2.3.4:27017/mydb?loadBalanced=true")

When using these options, if the specified server is not a load balancer, the client will fail all operations (until the server becomes a load balancer).

To treat the server as a load balancer even if it doesn't identify as such, use the connect: :load_balanced Ruby option or the connect=load_balanced URI option:

Mongo::Client.new([ '1.2.3.4:27017' ],
  database: 'mydb', load_balanced: true, connect: :load_balanced)

# Or using the URI syntax:
Mongo::Client.new("mongodb://1.2.3.4:27017/mydb?loadBalanced=true&connect=load_balanced")

MongoDB Atlas Connection

To connect to a MongoDB deployment on Atlas, first create a ::Mongo::Client instance using your cluster's connection string and other client options.

You can set the Stable API version as a client option to avoid breaking changes when you upgrade to a new server version.

The following code shows how you can specify the connection string and the Stable API client option when connecting to a MongoDB deployment and verify that the connection is successful:

require 'mongo'

# Replace the placeholders with your credentials
uri = "mongodb+srv://<username>:<password>@cluster0.sample.mongodb.net/?retryWrites=true&w=majority"

# Set the server_api field of the options object to Stable API version 1
options = { server_api: { version: "1" } }

# Create a new client and connect to the server
client = Mongo::Client.new(uri, options)

# Send a ping to confirm a successful connection
begin
  admin_client = client.use('admin')
  result = admin_client.database.command(ping: 1).documents.first
  puts "Pinged your deployment. You successfully connected to MongoDB!"
rescue Mongo::Error::OperationFailure => ex
  puts ex
ensure
  client.close
end

Connect to MongoDB Atlas from AWS Lambda

To learn how to connect to Atlas from AWS Lambda, see the Manage Connections with AWS Lambda documentation.

SRV URI Notes

When the driver connects to a mongodb+srv protocol <reference/connection-string/#dns-seedlist-connection-format> URI, keep in mind the following:

1. SRV URI lookup is performed synchronously when the client is constructed. If this lookup fails for any reason, client construction will fail with an exception. When a client is constructed with a list of hosts, the driver will attempt to contact and monitor those hosts for as long as the client object exists. If one of these hosts does not resolve initially but becomes resolvable later, the driver will be able to establish a connection to such a host when it becomes available. The initial SRV URI lookup must succeed on the first attempt; subsequent host lookups will be retried by the driver as needed.

2. The driver looks up URI options in the DNS TXT records corresponding to the SRV records. These options can be overridden by URI options specified in the URI and by Ruby options, in this order.

3. Because the URI options are retrieved in a separate DNS query from the SRV lookup, in environments with unreliable network connectivity the URI option query may fail when the SRV lookup succeeds. Such a failure would cause the driver to use the wrong auth source leading to authentication failures. This can be worked around by explicitly specifying the auth source:

   Mongo::Client.new("mongodb+srv://username:myRealPassword@cluster0.mongodb.net/test?w=majority&authSource=admin")

4. If the topology of the constructed Client object is unknown or a sharded cluster, the driver will begin monitoring the specified SRV DNS records for changes and will automatically update the list of servers in the cluster. The updates will stop if the topology becomes a single or a replica set.

Client Options

::Mongo::Client's constructor accepts a number of options configuring the behavior of the driver. The options can be provided in the options hash as Ruby options, in the URI as URI options, or both. If both a Ruby option and the analogous URI option are provided, the Ruby option takes precedence.

Ruby Options

{ read:
  { mode: :secondary,
    tag_sets: [ "data_center" => "berlin" ],
    max_staleness: 5,
  }
}

{ write_concern: { w: 2 } }

URI Options

Since the URI options are required to be in camel case, which is not the Ruby standard, the following table shows URI options and their corresponding Ruby options.

URI options are explained in detail in the Connection URI reference </reference/connection-string/>.

Timeout Options

server_selection_timeout

When executing an operation, the number of seconds to wait for the driver to find an appropriate server to send an operation to. Defaults to 30.

A value of 0 means no timeout.

When an invalid value (e.g. a negative value or a non-numeric value) is passed via the URI option, the invalid input is ignored with a warning. When an invalid value is passed directly to Client via a Ruby option, Client construction fails with an error.

In replica set deployments, this timeout should be set to exceed the typical replica set election times </core/replica-set-elections/> in order for the driver to transparently handle primary changes. This timeout also allows the application and the database to be started simultaneously; the application will wait up to this much time for the database to become available.

If the application server is behind a reverse proxy, server selection timeout should be lower than the request timeout configured on the reverse proxy (for example, this applies to deployments on Heroku which has a fixed 30 second timeout in the routing layer). In development this value can be lowered to provide quicker failure when the server is not running.

socket_timeout

The number of seconds to wait for a socket read or write to complete on regular (non-monitoring) connections. Default is no timeout.

A value of 0 means no timeout.

When an invalid value (e.g. a negative value or a non-numeric value) is passed via the URI option, the invalid input is ignored with a warning. When an invalid value is passed directly to Client via a Ruby option, Client construction fails with an error.

This timeout should take into account both network latency and operation duration. For example, setting this timeout to 5 seconds will abort queries taking more than 5 seconds to execute on the server with ::Mongo::Error::SocketTimeoutError.

Note that even though by default there is no socket timeout set, the operating system may still time out read operations depending on its configuration. The keepalive settings are intended to detect broken network connections (as opposed to aborting operations simply because they take a long time to execute).

Note that if an operation is timed out by the driver due to exceeding the socket_timeout value, it is not aborted on the server. For this reason it is recommended to use max_time_ms option for potentially long running operations, as this will abort their execution on the server.

This option does not apply to monitoring connections.

connect_timeout

The number of seconds to wait for a socket connection to be established to a server. Defaults to 10.

This timeout is also used as both connect timeout and socket timeout for monitoring connections.

When using a mongodb+srv:// URI, this timeout is also used for SRV and TXT DNS lookups. Note that the timeout applies per lookup; due to DNS suffix search lists, multiple lookups may be performed as part of a single name resolution.

wait_queue_timeout

The number of seconds to wait for a connection in the connection pool to become available. Defaults to 10.

As of driver version 2.11, this timeout should be set to a value at least as large as connect_timeout because connection pool now fully establishes connections prior to returning them, which may require several network round trips.

max_time_ms

Specified as an option on a particular operation, the number of milliseconds to allow the operation to execute for on the server. Not set by default.

Consider using this option instead of a socket_timeout for potentially long running operations to be interrupted on the server when they take too long.

wtimeout

The number of milliseconds to wait for a write to be acknowledged by the number of servers specified in the write concern. Not set by default, which instructs the server to apply its default. This option can be set globally on the client or passed to individual operations under :write_concern.

TLS Connections

To connect to the MongoDB deployment using TLS:

• Enable TLS connections in ::Mongo::Client.
• Specify the client TLS certificate.
• Specify the CA certificate to verify the server's TLS certificate.

TLS vs SSL Option Names

All MongoDB server versions supported by the Ruby driver (2.6 and higher) only implement TLS. 2.6 and higher servers do not use SSL.

For historical reasons, the Ruby option names pertaining to TLS configuration use the ssl rather than the tls prefix. The next major version of the Ruby driver (3.0) will use the tls prefix for Ruby option names.

The URI option names use the tls prefix, with one exception: there is a ssl URI option that is deprecated and equivalent to the tls URI option.

Enable TLS Connections

TLS must be explicitly requested on the client side when the deployment requires TLS connections - there is currently no automatic detection of whether the deployment requires TLS.

To request TLS connections, specify the following client options when constructing a ::Mongo::Client:

• The :ssl Ruby option.
• The tls URI option.
• The ssl URI option (deprecated).

Specify Client TLS Certificate

By default, MongoDB server will attempt to verify the connecting clients' TLS certificates, which requires the clients to specify their TLS certificates when connecting. This can be accomplished via:

• The :ssl_cert/:ssl_cert_object/:ssl_cert_string and :ssl_key/:ssl_key_object/:ssl_key_string/:ssl_key_pass_phrase Ruby options.
• The tlsCertificateKeyFile URI option.

When using the Ruby options, the client TLS certificate and the corresponding private key may be provided separately. For example, if the certificate is stored in client.crt and the private key is stored in client.key, a ::Mongo::Client may be constructed as follows:

client = Mongo::Client.new(["localhost:27017"],
  ssl: true,
  ssl_cert: 'path/to/client.crt',
  ssl_key: 'path/to/client.key',
  ssl_ca_cert: 'path/to/ca.crt',
)

ssl_cert, ssl_cert_string, ssl_key and ssl_key_string Ruby options also permit the certificate and the key to be provided in the same file or string, respectively. The files containing both certificate and private key frequently have the .pem extension. When both certificate and the private key are provided in the same file or string, both the certifcate and the key options must be utilized, as follows:

client = Mongo::Client.new(["localhost:27017"],
  ssl: true,
  ssl_cert: 'path/to/client.pem',
  ssl_key: 'path/to/client.pem',
  ssl_ca_cert: 'path/to/ca.crt',
)

When using the URI option, the certificate and the key must be stored in a file and both must be stored in the same file. Example usage:

client = Mongo::Client.new(
  "mongodb://localhost:27017/?tls=true&tlsCertificateKeyFile=path%2fto%2fclient.pem&tlsCertificateKeyFile=path%2fto%2fca.crt")

Modifying SSLContext [modifying-tls-context]

It may be desirable to further configure TLS options in the driver, for example by enabling or disabling certain ciphers. Currently, the Ruby driver does not provide a way to do this when initializing a ::Mongo::Client.

However, the Ruby driver provides a way to set global "TLS context hooks" --these are user-provided Procs that will be invoked before any TLS socket connection and can be used to modify the underlying OpenSSL::SSL::SSLContext object used by the socket.

To set the TLS context hooks, add Procs to the Mongo.tls_context_hooks array. This should be done before creating any Mongo::Client instances. For example, in a Rails application this code could be placed in an initializer.

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 the creation of every ::Mongo::Socket::SSL object.

It is possible to assign the entire array of hooks calling Mongo.tls_context_hooks=, but doing so will remove any previously assigned hooks. It is recommended to use the Array#push or Array#unshift methods to add new hooks.

It is also possible to remove hooks from Mongo.tls_context_hooks by storing a reference to the Procs somewhere else in the application, and then using Array#delete_if to remove the desired hooks.

Further information on configuring MongoDB server for TLS is available in the MongoDB manual </tutorial/configure-ssl/>.

Using Intermediate Certificates

It is possible to use certificate chains for both the client and the server certificates. When using chains, the certificate authority parameter should be configured to contain the trusted root certificates only; the intermediate certificates, if any, should be provided in the server or client certificates by concatenating them after the leaf server and client certificates, respectively.

:ssl_cert and :ssl_cert_string Ruby options, as well as tlsCertificateKeyFile URI option, support certificate chains. :ssl_cert_object Ruby option, which takes an instance of OpenSSL::X509::Certificate, does not support certificate chains.

The Ruby driver performs strict X.509 certificate verification, which requires that both of the following fields are set in the intermediate certificate(s):

• X509v3 Basic Constraints: CA: TRUE -- Can sign certificates
• X509v3 Key Usage: Key Cert Sign -- Can sign certificates

More information about these flags can be found in this Stack Overflow question.

It is a common pitfall to concatenate intermediate certificates to the root CA certificates passed in tlsCAFile / ssl_ca_cert options. By doing so, the intermediate certificates are elevated to trusted status and are themselves not verified against the actual CA root. More information on this issue is available in this mailing list post.

Specify CA Certificate

The driver will attempt to verify the server's TLS certificate by default, and will abort the connection if this verification fails. By default, the driver will use the default system root certificate store as the trust anchor. To specify the CA certificate that the server's certificate is signed with, use:

• The :ssl_ca_cert/:ssl_ca_cert_string/:ssl_ca_cert_object Ruby options
• The tlsCAFile URI option.

If any of these options are given, the server's certificate will be verified only against the specified CA certificate and the default system root certificate store will not be used.

To not perform server TLS certificate verification, which is not recommended, specify the ssl_verify: false Ruby option or the tlsInsecure=true URI option.

Specifying Multiple CA Certificates

The :ssl_ca_cert Ruby option and tlsCAFile URI option can be used with a file containing multiple certificates. All certificates thus referenced will become trust anchors.

The :ssl_ca_cert_object option takes an array of certificates, and thus can also be used to add multiple certificates as certificate authorities.

The :ssl_ca_cert_string option supports specifying only one CA certificate.

OCSP Verification

If the certificate provided by the server contains an OCSP endpoint URI, the driver will issue an OCSP request to the specified endpoint to verify the validity of the certificate.

The OCSP endpoint check may be disabled by setting the :ssl_verify_ocsp_endpoint Ruby option to false or by setting the tlsDisableOCSPEndpointCheck URI option to true when creating a client.

IPv4/IPv6 Connections

When a client is constructed with localhost as the host name, it will attempt an IPv4 connection only (i.e. if localhost resolves to 127.0.0.1 and ::1, the driver will only try to connect to 127.0.0.1).

When a client is constructed with hostnames other than localhost, it will attempt both IPv4 and IPv6 connections depending on the addresses that the hostnames resolve to. The driver respects the order in which getaddrinfo returns the addresses, and will attempt to connect to them sequentially. The first successful connection will be used.

The driver does not currently implement the Happy Eyeballs algorithm.

TCP Keepalive Configuration

Where allowed by system configuration and the Ruby language runtime, the driver enables TCP keepalive and, for each of the keepalive parameters listed below, sets the value of the respective parameter to the specified value if the system value can be determined and is higher than the listed driver value:

• tcp_keepalive_time: 120 seconds
• tcp_keepalive_intvl: 10 seconds
• tcp_keepalive_cnt: 9 probes

To use lower values, or to change the parameters in environments like JRuby that do not expose the required APIs, please adjust the parameters at the system level as described in the MongoDB Diagnostics FAQ keepalive section.

Connection Pooling

::Mongo::Client instances have a connection pool per server that the client is connected to. The pool creates connections on demand to support concurrent MongoDB operations issued by the application. There is no thread-affinity for connections.

The client instance opens one additional connection per known server for monitoring the server's state.

The size of each connection pool is capped at max_pool_size, which defaults to 5. When a thread in the application begins an operation on MongoDB, it tries to retrieve a connection from the pool to send that operation on. If there are some connections available in the pool, it checks out a connection from the pool and uses it for the operation. If there are no connections available and the size of the pool is less than the max_pool_size, a new connection will be created. If all connections are in use and the pool has reached its maximum size, the thread waits for a connection to be returned to the pool by another thread. If max_pool_size is set to zero, there is no limit for the maximum number of connections in the pool.

Each pool has a limit on the number of connections that can be concurrently connecting to a server. This limit is called max_connecting and defaults to

1. If the number of connections that are currently connecting to a server reaches this limit, the pool will wait for a connection attempt to succeed or fail before attempting to create a new connection. If your application has a large number of threads, you may want to increase max_connecting to avoid having threads wait for a connection to be established.

The number of seconds the thread will wait for a connection to become available is configurable. This setting, called wait_queue_timeout, is defined in seconds. If this timeout is reached, a Timeout::Error is raised. The default is 1 second.

As of driver version 2.11, the driver eagerly creates connections up to min_pool_size setting. Prior to driver version 2.11, the driver always created connections on demand. In all versions of the driver, once a connection is established, it will be kept in the pool by the driver as long as the pool size does not exceed min_pool_size.

Note that, if min_pool_size is set to a value greater than zero, the driver will establish that many connections to secondaries in replica set deployments even if the application does not perform secondary reads. The purpose of these connections is to provide faster failover when the primary changes.

Here is an example of estimating the number of connections a multi-threaded application will open: A client connected to a 3-node replica set opens 3 monitoring sockets. It also opens as many sockets as needed to support a multi-threaded application's concurrent operations on each server, up to max_pool_size. If the application only uses the primary (the default), then only the primary connection pool grows and the total connections is at most 8 (5 connections for the primary pool + 3 monitoring connections). If the application uses a read preference to query the secondaries, their pools also grow and the total connections can reach 18 (5 + 5 + 5 + 3).

The default configuration for a ::Mongo::Client works for most applications:

client = Mongo::Client.new(["localhost:27017"])

Create this client once for each process, and reuse it for all operations. It is a common mistake to create a new client for each request, which is very inefficient and not what the client was designed for.

To support extremely high numbers of concurrent MongoDB operations within one process, increase max_pool_size:

client = Mongo::Client.new(["localhost:27017"], max_pool_size: 200)

To support extremely high numbers of threads that share the same client within one process, increase max_connecting:

client = Mongo::Client.new(["localhost:27017"], max_pool_size: 200, max_connecting: 10)

Any number of threads are allowed to wait for connections to become available, and they can wait the default (1 second) or the wait_queue_timeout setting:

client = Mongo::Client.new(["localhost:27017"], wait_queue_timeout: 0.5)

When #close is called on a client by any thread, all connections are closed:

client.close

Note that when creating a client using the block syntax described above, the client is automatically closed after the block finishes executing.

Usage with Forking Servers [forking]

When using the Mongo Ruby driver in a Web application with a forking web server such as Puma, or when the application otherwise forks, each process (parent and child) must have its own client connections. This is because:

1. Background Ruby threads, such as those used by the Ruby MongoDB driver to monitor connection state, are not transferred to the child process.
2. File descriptors like network sockets are shared between parent and child processes, which can cause I/O conflicts.

Regarding (1), if you do not restart the driver's monitoring threads on the child process after forking, although your child may initially appear to function correctly, you will eventually see ::Mongo::Error::NoServerAvailable exceptions if/when your MongoDB cluster state changes, for example due to network errors or a maintenance event.

Regarding (2), if a child process reuses the parent's file descriptors, you will see ::Mongo::Error::SocketError errors with messages such as Errno::EPIPE: Broken pipe and EOFError: end of file reached.

When the Ruby driver is used in a web application, if possible, we recommend to not create any ::Mongo::Client instances in the parent process (prior to the workers being forked), and instead only create client instances in the workers.

Manually Handling Process Forks

Certain advanced use cases, such as [Puma's fork_worker option]), require ::Mongo::Client instances to be open in both the parent and child processes. In this case, you must handle client reconnection manually.

To do this, immediately before forking, close any existing client connections on your parent process. This will prevent the parent process from experiencing network and monitoring errors due to the child's reuse of the parent's file descriptors.

# Immediately before fork
client.close

Then, immediately after forking, reconnect your clients in the newly forked child process, which will respawn the driver's monitoring threads.

# Immediately after fork
client.reconnect

Most web servers provide hooks that can be used by applications to perform actions when the worker processes are forked. The recommended hooks are:

• For Puma, use before_fork and on_refork to close clients in the parent process and on_worker_boot to reconnect in the child processes.
• For Unicorn, before_fork to close clients in the parent process and after_fork to reconnect clients in the child processes.
• For Passenger, starting_worker_process to reconnect clients in the child processes (Passenger does not appear to have a pre-fork hook).

Refer to Mongoid's "Usage with Forking Servers" documentation for further examples.

Troubleshooting

The client's summary method returns the current state of the client, including servers that the client is monitoring and their state. If any of the servers are not being monitored, this is indicated by the NO-MONITORING flag.

A normally operating client will produce a summary similar to the following:

client.summary
=> "#<Client cluster=#<Cluster
  topology=ReplicaSetWithPrimary[localhost:14420,localhost:14421,localhost:14422,localhost:14423,name=ruby-driver-rs,v=1,e=7fffffff000000000000001f]
  servers=[#<Server address=localhost:14420 PRIMARY replica_set=ruby-driver-rs pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14421 SECONDARY replica_set=ruby-driver-rs pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14422 SECONDARY replica_set=ruby-driver-rs pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14423 ARBITER replica_set=ruby-driver-rs>]>>"

A client that is missing background threads will produce a summary similar to the following:

client.summary
=> "#<Client cluster=#<Cluster
  topology=ReplicaSetWithPrimary[localhost:14420,localhost:14421,localhost:14422,localhost:14423,name=ruby-driver-rs,v=1,e=7fffffff000000000000001f]
  servers=[#<Server address=localhost:14420 PRIMARY replica_set=ruby-driver-rs NO-MONITORING pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14421 SECONDARY replica_set=ruby-driver-rs NO-MONITORING pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14422 SECONDARY replica_set=ruby-driver-rs NO-MONITORING pool=#<ConnectionPool size=0 (0-5) used=0 avail=0 pending=0>>,
    #<Server address=localhost:14423 ARBITER replica_set=ruby-driver-rs>]>>"

Retryable Reads

The driver implements two mechanisms for retrying reads: modern and legacy. As of driver version 2.9.0, the modern mechanism is used by default, and the legacy mechanism is deprecated.

Modern Retryable Reads

When the modern mechanism is used, read operations are retried once in the event of a network error, a "not master" error, or a "node is recovering" error. The following operations are covered:

• Collection#find and related methods
• Collection#aggregate
• Collection#count, Collection#count_documents
• Change stream helpers: Collection#watch, Database#watch, Client#watch
• Enumeration commands: Client#list_mongo_databases, Client#list_databases, Client#database_names, Database#collection_names, Database#collections, Database#list_collections, Collection#indexes

When an operation returns a cursor, only the initial read command can be retried. getMore operations on cursors are not retried by driver version 2.9.0 or newer. Additionally, when a read operation is retried, a new server for the operation is selected; this may result in the retry being sent to a different server from the one which received the first read.

The behavior of modern retryable reads is covered in detail by the retryable reads specification.

Note that the modern retryable reads can only be used with MongoDB 3.6 and higher servers. When used with MongoDB 3.4 and lower servers, Ruby driver version 2.9.0 and higher will not retry reads by default - the application must explicitly request legacy retryable reads by setting the retry_reads: false client option or using retryReads=false URI option.

Legacy Retryable Reads

The legacy read retry behavior of the Ruby driver is available by setting the retry_reads: false client option or passing the retryReads=false URI option to the client.

When using legacy read retry behavior, the number of retries can be set by specifying the max_read_retries client option. When using driver version 2.9.0 or higher, the set of operations which would be retried with legacy retryable reads is identical to the one described above for modern retryable reads. In older driver versions the behavior of legacy retryable writes was different in that some of the operations were not retried.

As of driver version 2.9.0, legacy read retries perform server selection prior to retrying the operation, as modern retriable writes do. In older driver versions read retries would be sent to the same server which the initial read was sent to.

Disabling Retryable Reads

To disable all read retries, set the following client options: retry_reads: false, max_read_retries: 0.

Retryable Writes

The driver implements two mechanisms for retrying writes: modern and legacy. As of driver version 2.9.0, the modern mechanism is used by default on servers that support it, and the legacy mechanism is deprecated and disabled by default on all server versions.

The following write methods used in day-to-day operations on collections are subject to write retries:

• collection#insert_one
• collection#update_one
• collection#delete_one
• collection#replace_one
• collection#find_one_and_update
• collection#find_one_and_replace
• collection#find_one_and_delete
• collection#bulk_write (for all single statement ops, i.e. not for update_many or delete_many)

Modern Retryable Writes

The modern mechanism will retry failing writes once when the driver is connected to a MongoDB 3.6 or higher replica set or a sharded cluster, because they require an oplog on the serer. Modern mechanism will not retry writes when the driver is connected to a standalone MongoDB server or server versions 3.4 or older.

The following errors will cause writes to be retried:

• Network errors including timeouts
• "not master" errors
• "node is recovering" errors

Prior to retrying the write the driver will perform server selection, since the server that the original write was sent to is likely no longer usable.

Legacy Retryable Writes

If modern retryable writes mechanism is disabled by setting the client option retry_writes: false or by using the retryWrites=false URI option, the driver will utilize the legacy retryable writes mechanism. The legacy mechanism retries writes on the same operations as the modern mechanism. By default the legacy mechanism retries once, like the modern mechanism does; to change the number of retries, set :max_write_retries client option.

The difference between legacy and modern retry mechanisms is that the legacy mechanism retries writes for a different set of errors compared to the modern mechanism, and specifically does not retry writes when a network timeout is encountered.

Disabling Retryable Writes

To disable all write retries, set the following client options: retry_writes: false, max_write_retries: 0.

Logging

You can either use the default global driver logger or set your own. To set your own:

Mongo::Logger.logger = other_logger

See the Ruby Logger documentation for more information on the default logger API and available levels.

Changing the Logger Level

To change the logger level:

Mongo::Logger.logger.level = Logger::WARN

For more control, a logger can be passed to a client for per-client control over logging.

my_logger = Logger.new(STDOUT)
Mongo::Client.new([ '127.0.0.1:27017' ], :database => 'test', :logger => my_logger )

Truncation

The default logging truncates logs at 250 characters by default. To turn this off pass an option to the client instance.

Mongo::Client.new([ '127.0.0.1:27017' ], :database => 'test', :truncate_logs => false )

Compression

To use wire protocol compression, at least one compressor must be explicitly requested using either the :compressors Ruby option or the compressors URI option. If no compressors are explicitly requested, the driver will not use compression, even if the required dependencies for one or more compressors are present on the system.

The driver chooses the first compressor of the ones requested that is also supported by the server. The driver currently supports zstd, snappy and zlib compressors. zstd compressor is recommended as it produces the highest compression at the same CPU consumption compared to the other compressors. For maximum server compatibility all three compressors can be specified, e.g. as compressors: ["zstd", "snappy", "zlib"].

zstd compressor requires the zstd-ruby library to be installed. snappy compressor requires the snappy library to be installed. If zstd or snappy compression is requested, and the respective library is not loadable, the driver will raise an error during ::Mongo::Client creation. zlib compression requires the zlib standard library extension to be present.

The server support for various compressors is as follows:

• zstd requires and is enabled by default in MongoDB 4.2 or higher.
• snappy requires MongoDB 3.4 or higher and is enabled by default in MongoDB 3.6 or higher.
• zlib requires MongoDB 3.6 or higher and is enabled by default in MongoDB 4.2 and higher.

Server API Parameters

Starting with MongoDB 5.0, applications can request that the server behaves in accordance with a particular server API version.

Server API parameters can be specified via the :server_api option to Client. These parameters cannot be provided via a URI.

Currently the only defined API version is "1". It can be requested as follows:

client = Mongo::Client.new(['localhost'], server_api: {version: "1"})

MongoDB server defines API versions as string values. For convenience, if the API version is provided as an integer, the Ruby driver will stringify it and send it to the server as a string:

client = Mongo::Client.new(['localhost'], server_api: {version: 1})

Note that the server may define API versions that are not stringified integers. Applications must not assume that all legal API versions can be expressed as integers.

When a particular API version is requested, operations which are part of that API version behave as specified in that API version. Operations which are not part of the specified API version behave as they would had the API version not been specified at all. Operations whose behavior is subject to the configured API version are commands including command arguments, queries, aggregation pipeline stages and arguments.

Applications may request that the server rejects all operations which are not part of the specified API version by setting the :strict option:

client = Mongo::Client.new(['localhost'], server_api: {version: "1", strict: true})

For example, since the :tailable option is not part of the server API version 1, the following query would fail:

client = Mongo::Client.new(['localhost'], server_api: {version: "1", strict: true})
client['collection'].find({}, tailable: true)
# => Mongo::Error::OperationFailure (BSON field 'FindCommand.tailable' is not allowed with apiStrict:true. (323) (on localhost:27017, modern retry, attempt 1))

Applications may request that the server rejects all operations which are deprecated in the specified API version by setting the :deprecation_errors option:

client = Mongo::Client.new(['localhost'], server_api: {version: "1", deprecation_errors: true})

Note that, as of this writing, there are no deprecated operations in API version "1".

If the server API parameters have been defined on a Client object, they will be sent by the client as part of each[^1] executed operation.

MongoDB servers prior to 5.0 do not recognize the API parameters, and will produce a variety of errors should the application configure them. The Ruby driver will send the API parameters to all MongoDB 3.6 and newer servers, but the API parameters should only be configured when the application is communicating with MongoDB 5.0 or newer servers. The API parameters cannot be sent to MongoDB 3.4 and older servers that use the legacy wire protocol; if an application configures the API parameters and connects to MongoDB 3.4 or older servers, the driver will produce an error on every operation.

The command helper <arbitrary-commands> permits the application to send manually constructed commands to the server. If the client is not configured with server API parameters, the command helper may be used to issue commands with API parameters:

client.database.command(
  ping: 1,
  apiVersion: "1",
  apiStrict: false,
  apiDeprecationErrors: false,
)

If the client is configured with server API parameters, the command helper may not be used to issue commands with server API parameters. This includes the case when the server API parameters provided to the client and to the command helper are identical. If a client is constructed with server API parameters, to send different API parameters (or none at all) a new client must be constructed, either from scratch or using the with method.

The server API parameters may only be specified on the client level. They may not be specified on the database, collection, session, transaction or individual operation level.

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.

Production Configuration

Please consider the following when deploying an application using the Ruby driver in production:

• As of driver version 2.11, the :min_pool_size client option is completely respected - the driver will create that many connections to each server identified as a standalone, primary or secondary. In previous driver versions the driver created connections on demand. Applications using :min_pool_size will see an increase in the number of idle connections to all servers as of driver version 2.11, and especially to secondaries in replica set deployments and to nodes in sharded clusters.
• If the application is reverse proxied to by another web server or a load balancer, server_selection_timeout should generally be set to a lower value than the reverse proxy's read timeout. For exampe, Heroku request timeout is 30 seconds and is not configurable; if deploying a Ruby application using MongoDB to Heroku, consider lowering server selection timeout to 20 or 15 seconds.

Feature Flags

The following is a list of feature flags that the Mongo Ruby Driver provides:

the documents included in that view, instead of all documents in the        
collection. When this flag is on, the view fiter is ignored and the         
aggregation is applied over all of the documents in the view's              
collection. (default: true)                                                 |

| broken_view_options | When this flag is turned off, the view options will be correctly
propagated to the aggregate, count, count_documents,
distinct, and estimated_document_count mehods. When this flag is
switched on, the view options will be ignored in those methods.
(default: true) | | validate_update_replace | 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. (default: false) |

These feature flags can be set directly on the Mongo module or using the options method:

Mongo.validate_update_replace = true
Mongo.options = { validate_update_replace: true }

[^1]: getMore commands and commands in transactions do not accept API parameters, thus the driver will not send them in these cases.