CRUD Operations on Replica Sets

Overview

In this guide, you can learn how to use the write concern, read concern, and read preference configurations to modify the way that MongoDB runs create, read, update, and delete (CRUD) operations on replica sets.

You can set write concern, read concern, and read preference options at the following levels:

• Client, which sets the default for all operation executions unless overridden

• Session

• Transaction

• Database

• Collection

The preceding list also indicates the increasing order of precedence of the option settings. For example, if you set a read concern level for a transaction, it will override a read concern level set for the client.

These options allow you to customize the causal consistency and availability of the data in your replica sets.

Write Concern

The write concern specifies the level of acknowledgement requested from MongoDB for write operations, such as an insert or update, before the operation successfully returns. Operations that do not specify an explicit write concern inherit the global default write concern settings.

For more information, see Write Concern in the MongoDB Server manual. For detailed API documentation, see the Write Concern API documentation.

The following table describes the write_concern parameters:

client = Mongo::Client.new(['IP_ADDRESS_001:27017'], database: 'myDB')
myDB = client.database
myCollection = myDB[:myCollection]
myCollection.insert_one(
  { name: 'anotherDocumentName' },
  write: {
    w: 2,
    wtimeout: 5000
  }
)

myDoc = { name: 'New Document' }
new_write_concern = Mongo::WriteConcern.get(myDB.write_concern)
myDB[:myCollection].with(write: new_write_concern).insert_one(myDoc)

Read Concern

The read concern specifies the following behaviors:

• Level of causal consistency across replica sets

• Isolation guarantees maintained during a query

You can specify the read concern setting by using the level parameter. The default read concern level is local. This means that the client returns the data from the replica set member that the client is connected to, with no guarantee that the data has been written to all replica set members.

For more information about read concerns or read concern levels, see Read Concern in the MongoDB Server manual. For more detail on the read_concern type and definitions of the read concern levels, see Read Concern in the API documentation.

pipeline = [
  { "$match" => { category: 'KITCHENWARE' } },
  { "$unset" => ['_id', 'category'] }
]
result = myCollection.aggregate(pipeline,
read: { read_concern: { level: :available } })

client = Mongo::Client.new(['IP_ADDRESS_001:27017'], 
  database: 'mydb', 
  read_concern: { level: :local })
myDB = client.database

Read Preference

The read preference determines which member of a replica set MongoDB reads when running a query.

For more detailed API documentation, see the Read Preference API documentation.

The following table shows options you can use to customize how the server evaluates members:

transaction_options = {
  read: { mode: :primary },
  read_concern: { level: :local },
  write_concern: { w: :majority }
}
session = client.start_session
session.start_transaction(transaction_options)
session.commit_transaction
# ...
rescue => e
    session.abort_transaction
    puts "Transaction aborted due to an error: #{e.message}"
ensure
    session.end_session
end

uri = 'mongodb+srv://<user>:<password>@<cluster-url>'
options = {
  read: {
    mode: :secondary,
    max_staleness: 120
  }
}
client = Mongo::Client.new(uri, options)
myDB = client.database

API Documentation

To learn more about the methods and types mentioned in this guide, see the following API documentation:

• Write Concern

• Read Concern

• Read Preference