Docs Menu
Docs Home
MongoDB Manual
/ / / /


On this page

  • Definition
  • Behavior
  • Example

Starts a multi-document transaction associated with the session. At any given time, you can have at most one open transaction for a session.


This operation is a no-op. The transaction won't start on the server until the first command is sent on the session. Therefore, the snapshot time of the transaction won't be set until the first command is sent on the session.

Multi-document transactions are available for both sharded clusters and replica sets.


Within a transaction, you can only specify read and write (CRUD) operations on existing collections. For example, a multi-document transaction cannot include an insert operation that would result in the creation of a new collection.

The Session.startTransaction() method can take a document with the following options:

{ readConcern: { level: <level>}, writeConcern: { w: <value>, j: <boolean>, wtimeout: <number> } }

Optional. A document that specifies the read concern for all operations in the transaction, overriding operation-specific read concern.

You can specify one of the following read concern levels:

For "local" and "majority" read concern, MongoDB may sometimes substitute a stronger read concern.


Optional. A document that specifies the write concern for the transaction. This write concern applies to the transaction commit and abort operations.

The operations within the transaction use "w: 1", overriding operation-specific write concern.

If you commit using "w: 1" write concern, your transaction can be rolled back during the failover process.

For MongoDB Drivers, transactions use the client-level write concern as the default.


If running with access control, you must have privileges for the operations in the transaction.

For multi-document transactions:

  • You can create collections and indexes in transactions. For details, see Create Collections and Indexes in a Transaction

  • The collections used in a transaction can be in different databases.


    You cannot create new collections in cross-shard write transactions. For example, if you write to an existing collection in one shard and implicitly create a collection in a different shard, MongoDB cannot perform both operations in the same transaction.

  • You cannot write to capped collections.

  • You cannot use read concern "snapshot" when reading from a capped collection. (Starting in MongoDB 5.0)

  • You cannot read/write to collections in the config, admin, or local databases.

  • You cannot write to system.* collections.

  • You cannot return the supported operation's query plan using explain or similar commands.

  • For cursors created outside of a transaction, you cannot call getMore inside the transaction.

  • For cursors created in a transaction, you cannot call getMore outside the transaction.


Excluding the following query operator expressions:

The method uses the $match aggregation stage for the query and $group aggregation stage with a $sum expression to perform the count.

Available on unsharded collections.

For sharded collections, use the aggregation pipeline with the $group stage. See Distinct Operation.

If the update or replace operation is run with upsert: true on a non-existing collection, the collection is implicitly created.


See also:

If run on a non-existing collection, the collection is implicitly created.

If run on a non-existing collection, the collection is implicitly created.

If run on a non-existing collection, the collection is implicitly created.

Operations that affect the database catalog, such as creating or dropping a collection or an index, are not allowed in multi-document transactions. For example, a multi-document transaction cannot include an insert operation that would result in the creation of a new collection. See Restricted Operations.

Informational commands, such as hello, buildInfo, connectionStatus (and their helper methods) are allowed in transactions; however, they cannot be the first operation in the transaction.

Transactions support read preference primary.

While the transaction is open, no data changes made by operations in the transaction is visible outside the transaction:

  • When a transaction commits, all data changes made in the transaction are saved and visible outside the transaction. That is, a transaction will not commit some of its changes while rolling back others.

    Until a transaction commits, the data changes made in the transaction are not visible outside the transaction.

    However, when a transaction writes to multiple shards, not all outside read operations need to wait for the result of the committed transaction to be visible across the shards. For example, if a transaction is committed and write 1 is visible on shard A but write 2 is not yet visible on shard B, an outside read at read concern "local" can read the results of write 1 without seeing write 2.

  • When a transaction aborts, all data changes made by the writes in the transaction are discarded without ever becoming visible and the transaction ends.

Consider a scenario where as changes are made to an employee's record in the hr database, you want to ensure that the events collection in the reporting database are in sync with the hr changes. That is, you want to ensure that these writes are done as a single transaction, such that either both operations succeed or fail.

The employees collection in the hr database has the following documents:

{ "_id" : ObjectId("5af0776263426f87dd69319a"), "employee" : 3, "name" : { "title" : "Mr.", "name" : "Iba Ochs" }, "status" : "Active", "department" : "ABC" }
{ "_id" : ObjectId("5af0776263426f87dd693198"), "employee" : 1, "name" : { "title" : "Miss", "name" : "Ann Thrope" }, "status" : "Active", "department" : "ABC" }
{ "_id" : ObjectId("5af0776263426f87dd693199"), "employee" : 2, "name" : { "title" : "Mrs.", "name" : "Eppie Delta" }, "status" : "Active", "department" : "XYZ" }

The events collection in the reporting database has the following documents:

{ "_id" : ObjectId("5af07daa051d92f02462644a"), "employee" : 1, "status" : { "new" : "Active", "old" : null }, "department" : { "new" : "ABC", "old" : null } }
{ "_id" : ObjectId("5af07daa051d92f02462644b"), "employee" : 2, "status" : { "new" : "Active", "old" : null }, "department" : { "new" : "XYZ", "old" : null } }
{ "_id" : ObjectId("5af07daa051d92f02462644c"), "employee" : 3, "status" : { "new" : "Active", "old" : null }, "department" : { "new" : "ABC", "old" : null } }

The following example opens a transaction, updates an employee's status to Inactive in the employees status and inserts a corresponding document to the events collection, and commits the two operations as a single transaction.

// Runs the txnFunc and retries if TransientTransactionError encountered
function runTransactionWithRetry(txnFunc, session) {
while (true) {
try {
txnFunc(session); // performs transaction
} catch (error) {
// If transient error, retry the whole transaction
if (error?.errorLabels?.includes("TransientTransactionError") ) {
print("TransientTransactionError, retrying transaction ...");
} else {
throw error;
// Retries commit if UnknownTransactionCommitResult encountered
function commitWithRetry(session) {
while (true) {
try {
session.commitTransaction(); // Uses write concern set at transaction start.
print("Transaction committed.");
} catch (error) {
// Can retry commit
if (error?.errorLabels?.includes("UnknownTransactionCommitResult") ) {
print("UnknownTransactionCommitResult, retrying commit operation ...");
} else {
print("Error during commit ...");
throw error;
// Updates two collections in a transactions
function updateEmployeeInfo(session) {
employeesCollection = session.getDatabase("hr").employees;
eventsCollection = session.getDatabase("reporting").events;
session.startTransaction( { readConcern: { level: "snapshot" }, writeConcern: { w: "majority" } } );
employeesCollection.updateOne( { employee: 3 }, { $set: { status: "Inactive" } } );
eventsCollection.insertOne( { employee: 3, status: { new: "Inactive", old: "Active" } } );
} catch (error) {
print("Caught exception during transaction, aborting.");
throw error;
// Start a session.
session = db.getMongo().startSession( { readPreference: { mode: "primary" } } );
runTransactionWithRetry(updateEmployeeInfo, session);
} catch (error) {
// Do something with error
} finally {
← Session.commitTransaction()