Navigation
This is an upcoming (i.e. in progress) version of the manual.

Retryable Writes

New in version 3.6.

Retryable writes allow MongoDB drivers to automatically retry certain write operations a single time if they encounter network errors, or if they cannot find a healthy primary in the replica sets or sharded cluster. [1]

Prerequisites

Retryable writes have the following requirements:

Supported Deployment Topologies
Retryable writes require a replica set or sharded cluster, and do not support standalone instances.
Supported Storage Engine
Retryable writes require a storage engine supporting document-level locking, such as the WiredTiger or in-memory storage engines.
3.6+ MongoDB Drivers

Clients require MongoDB drivers updated for MongoDB 3.6 or greater:

Java 3.6+

Python 3.6+

C 1.9+

C# 2.5+

Node 3.0+

Ruby 2.5+

Perl 2.0+

PHPC 1.4+

Scala 2.2+

MongoDB Version
The MongoDB version of every node in the cluster must be 3.6 or greater, and the featureCompatibilityVersion of each node in the cluster must be 3.6 or greater. See setFeatureCompatibilityVersion for more information on the featureCompatibilityVersion flag.
Write Acknowledgment
Write operations issued with a Write Concern of 0 are not retryable.

Retryable Writes and Multi-Document Transactions

New in version 4.0.

The transaction commit and abort operations are retryable write operations. If the commit operation or the abort operation encounters an error, MongoDB drivers retry the operation a single time regardless of whether retryWrites is set to true.

The write operations inside the transaction are not individually retryable, regardless of whether retryWrites is set to true.

For more information on transactions, see Transactions.

Enabling Retryable Writes

MongoDB drivers do not enable retryable writes by default with the exception of transaction commit and abort operations.

MongoDB Drivers

To enable retryable writes in MongoDB drivers, add the retryWrites option to your connection string:

mongodb://localhost/?retryWrites=true
mongo shell

To enable retryable writes in the mongo shell, use the --retryWrites command line option:

mongo --retryWrites

Retryable Write Operations

The following write operations are retryable when issued with acknowledged write concern; e.g., Write Concern cannot be {w: 0}.

Note

The write operations inside the transactions are not individually retryable.

Methods Descriptions
Insert operations.
Single-document update operations. [1]
Single document delete operations.
findAndModify operations. All findAndModify operations are single document operations.

db.collection.bulkWrite() with the following write operations:

Bulk write operations that only consist of the single-document write operations. A retryable bulk operation can include any combination of the specified write operations but cannot include any multi-document write operations, such as updateMany.

Bulk operations for:

Bulk write operations that only consist of the single-document write operations. A retryable bulk operation can include any combination of the specified write operations but cannot include any multi-document write operations, such as update which specifies true for the multi option.

Updates to Shard Key Values

Starting in MongoDB 4.2, you can update a document’s shard key value (unless the shard key field is the immutable _id field) by issuing single-document update/findAndModify operations either as a retryable write or in a transaction. For details, see Change a Document’s Shard Key Value.

[1](1, 2)

MongoDB 4.2 will retry certain single-document upserts (update with upsert: true and multi: false) that encounter a duplicate key exception. See Duplicate Key Errors on Upsert for conditions.

Prior to MongoDB 4.2, MongoDB would not retry upsert operations that encountered a duplicate key error.

Behavior

Persistent Network Errors

MongoDB retryable writes make only one retry attempt. This helps address transient network errors and replica set elections, but not persistent network errors.

Failover Period

If the driver cannot find a healthy primary in the destination replica set or sharded cluster shard, the drivers wait serverSelectionTimeoutMS milliseconds to determine the new primary before retrying. Retryable writes do not address instances where the failover period exceeds serverSelectionTimeoutMS.

Warning

If the client application becomes temporarily unresponsive for more than the localLogicalSessionTimeoutMinutes after issuing a write operation, there is a chance that when the client applications starts responding (without a restart), the write operation may be retried and applied again.

Duplicate Key Errors on Upsert

MongoDB 4.2 will retry single-document upsert operations (i.e upsert : true and multi : false) that fail due to a duplicate key error only if the operation meets all of the following conditions:

  • The target collection has a unique index that caused the duplicate key error.

  • The update match condition is either:

    • A single equality predicate

      { "fieldA" : "valueA" },

      or

    • a logical AND of equality predicates

      { "fieldA" : "valueA", "fieldB" : "valueB" }

  • The set of fields in the unique index key pattern matches the set of fields in the update query predicate.

  • The update operation does not modify any of the fields in the query predicate.

The following table contains examples of upsert operations that the server can or cannot retry on a duplicate key error:

Unique Index Key Pattern Update Operation Retryable
{ _id : 1 }
db.collName.updateOne(
  { _id : ObjectId("1aa1c1efb123f14aaa167aaa") },
  { $set : { fieldA : 25 } },
  { upsert : true }
)
Yes
{ fieldA : 1 }
db.collName.updateOne(
  { fieldA : { $in : [ 25 ] } },
  { $set : { fieldB : "someValue" } },
  { upsert : true }
)
Yes
{
  fieldA : 1,
  fieldB : 1
}
db.collName.updateOne(
  { fieldA : 25, fieldB : "someValue" },
  { $set : { fieldC : false } },
  { upsert : true }
)
Yes
{ fieldA : 1 }
db.collName.updateOne(
  { fieldA : { $lte : 25 } },
  { $set : { fieldC : true } },
  { upsert : true }
)

No

The query predicate on fieldA is not an equality

{ fieldA : 1 }
db.collName.updateOne(
  { fieldA : { $in : [ 25 ] } },
  { $set : { fieldA : 20 } },
  { upsert : true }
)

No

The update operation modifies fields specified in the query predicate.

{ _id : 1 }
db.collName.updateOne(
  { fieldA : { $in : [ 25 ] } },
  { $set : { fieldA : 20 } },
  { upsert : true }
)

No

The set of query predicate fields (fieldA) does not match the set of index key fields (_id).

{ fieldA : 1 }
db.collName.updateOne(
  { fieldA : 25, fieldC : true },
  { $set : { fieldD : false } },
  { upsert : true }
)

No

The set of query predicate fields (fieldA, fieldC) does not match the set of index key fields (fieldA).

Prior to MongoDB 4.2, MongoDB retryable writes did not support retrying upserts which failed due to duplicate key errors.

Diagnostics

New in version 3.6.3.

The serverStatus command, and its mongo shell helper db.serverStatus() includes statistics on retryable writes in the transactions section.