Version 10.0.2

ACID properties

Transactions provide a powerful abstraction for multiple threads to operate on data concurrently because they have the following properties:

  • Atomicity: all or none of a transaction is completed.
  • Consistency: if each transaction maintains some property when considered separately, then the combined effect of executing the transactions concurrently will maintain the same property.
  • Isolation: developers can reason about transactions as if they run single-threaded.
  • Durability: once a transaction commits, its updates cannot be lost.

WiredTiger supports transactions with the following caveats to the ACID properties:

  • the maximum level of isolation supported is snapshot isolation and all updates must be done at snapshot isolation. See Isolation levels for more details.
  • transactional updates are made durable by a combination of checkpoints and logging. See Checkpoint durability for information on checkpoint durability and Commit-level durability for information on commit-level durability.

Transactional API

In WiredTiger, transaction operations are methods off the WT_SESSION class.

Applications call WT_SESSION::begin_transaction to start a new transaction. Operations subsequently performed using that WT_SESSION handle, including operations on any cursors open in that WT_SESSION handle (whether opened before or after the WT_SESSION::begin_transaction call), are part of the transaction and their effects committed by calling WT_SESSION::commit_transaction, or discarded by calling WT_SESSION::rollback_transaction. Applications that use Application-specified Transaction Timestamps can utilize the WT_SESSION::prepare_transaction API as a basis for implementing a two phase commit protocol.

If WT_SESSION::commit_transaction returns an error for any reason, the transaction was rolled back, not committed.

When transactions are used, data operations can encounter a conflict and fail with the WT_ROLLBACK error. If this error occurs, transactions should be rolled back with WT_SESSION::rollback_transaction and the operation retried.

The WT_SESSION::rollback_transaction method implicitly resets all cursors in the session as if the WT_CURSOR::reset method was called, discarding any cursor position as well as any key and value.

* Cursors may be opened before or after the transaction begins, and in either case, subsequent
* operations are included in the transaction. Opening cursors before the transaction begins
* allows applications to cache cursors and use them for multiple operations.
error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
error_check(session->begin_transaction(session, NULL));
cursor->set_key(cursor, "key");
cursor->set_value(cursor, "value");
switch (cursor->update(cursor)) {
case 0: /* Update success */
error_check(session->commit_transaction(session, NULL));
* If commit_transaction succeeds, cursors remain positioned; if commit_transaction fails,
* the transaction was rolled-back and all cursors are reset.
case WT_ROLLBACK: /* Update conflict */
default: /* Other error */
error_check(session->rollback_transaction(session, NULL));
/* The rollback_transaction call resets all cursors. */
* Cursors remain open and may be used for multiple transactions.

Applications can call WT_SESSION::reset_snapshot to reset snapshots for snapshot isolation transactions to update their existing snapshot. It raises an error when this API is used for isolation other than snapshot isolation mode or when the session has performed any write operations. This API internally releases the current snapshot and gets the new running transactions snapshot to avoid pinning the content in the database that is no longer needed. Applications that don't use read_timestamp for the search may see different results compared to earlier with the updated snapshot.

* Resets snapshots for snapshot isolation transactions to update their existing snapshot.
* It raises an error when this API is used for isolation other than snapshot isolation
* mode.
error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
error_check(session->begin_transaction(session, "isolation=snapshot"));
cursor->set_key(cursor, "some-key");
error_check(session->commit_transaction(session, NULL));

Implicit transactions

If a cursor is used when no explicit transaction is active in a session, reads are performed at the isolation level of the session, set with the isolation key to WT_CONNECTION::open_session, and successful updates are automatically committed before the update operation returns.

Any operation consisting of multiple related updates should be enclosed in an explicit transaction to ensure the updates are applied atomically.

If an implicit transaction successfully commits, the cursors in the WT_SESSION remain positioned. If an implicit transaction fails, all cursors in the WT_SESSION are reset, as if WT_CURSOR::reset were called, discarding any position or key/value information they may have.

See Cursors and Transactions for more information.

Concurrency control

WiredTiger uses optimistic concurrency control algorithms. This avoids the bottleneck of a centralized lock manager and ensures transactional operations do not block: reads do not block writes, and vice versa.

Further, writes do not block writes, although concurrent transactions updating the same value will fail with WT_ROLLBACK. Some applications may benefit from application-level synchronization to avoid repeated attempts to rollback and update the same value.

Operations in transactions may also fail with the WT_ROLLBACK error if some resource cannot be allocated after repeated attempts. For example, if the cache is not large enough to hold the updates required to satisfy transactional readers, an operation may fail and return WT_ROLLBACK.

Isolation levels

WiredTiger supports read-uncommitted, read-committed and snapshot isolation levels; the default isolation level is snapshot, and all updates must be done at snapshot isolation.

  • read-uncommitted: Transactions can see changes made by other transactions before those transactions are committed. Dirty reads, non-repeatable reads and phantoms are possible.
  • read-committed: Transactions cannot see changes made by other transactions before those transactions are committed. Dirty reads are not possible; non-repeatable reads and phantoms are possible. Committed changes from concurrent transactions become visible when no cursor is positioned in the read-committed transaction.
  • snapshot: Transactions read the versions of records committed before the transaction started. Dirty reads and non-repeatable reads are not possible; phantoms are possible.

    Snapshot isolation is a strong guarantee, but not equivalent to a single-threaded execution of the transactions, known as serializable isolation. Concurrent transactions T1 and T2 running under snapshot isolation may both commit and produce a state that neither (T1 followed by T2) nor (T2 followed by T1) could have produced, if there is overlap between T1's reads and T2's writes, and between T1's writes and T2's reads.

The transaction isolation level can be configured on a per-transaction basis:

/* A single transaction configured for snapshot isolation. */
error_check(session->open_cursor(session, "table:mytable", NULL, NULL, &cursor));
error_check(session->begin_transaction(session, "isolation=snapshot"));
cursor->set_key(cursor, "some-key");
cursor->set_value(cursor, "some-value");
error_check(session->commit_transaction(session, NULL));

Additionally, the default transaction isolation can be configured and re-configured on a per-session basis:

/* Open a session configured for read-uncommitted isolation. */
error_check(conn->open_session(conn, NULL, "isolation=read-uncommitted", &session));
/* Re-configure a session for snapshot isolation. */
error_check(session->reconfigure(session, "isolation=snapshot"));

Application-specified Transaction Timestamps

Timestamp overview

Some applications have their own notion of time, including an expected commit order for transactions that may be inconsistent with the order assigned by WiredTiger. We assume applications can represent their notion of a timestamp as an unsigned 64-bit integral value that generally increases over time. For example, a counter could be incremented to generate transaction timestamps, if that is sufficient for the application.

Applications can assign explicit commit timestamps to transactions, then read "as of" a timestamp. The timestamp mechanism operates in parallel with WiredTiger's internal transaction ID management. It is recommended that once timestamps are in use for a particular table, all subsequent updates also use timestamps.

Using transactions with timestamps

Applications that use timestamps will generally provide a timestamp at WT_SESSION::transaction_commit that will be assigned to all updates that are part of the transaction. WiredTiger also provides the ability to set a different commit timestamp for different updates in a single transaction. This can be done by calling WT_SESSION::timestamp_transaction repeatedly to set a new commit timestamp between a set of updates for the current transaction. This gives the ability to commit updates with different read "as of" timestamps in a single transaction.

Setting a read timestamp in WT_SESSION::begin_transaction forces a transaction to run at snapshot isolation and ignore any commits with a newer timestamp.

Commit timestamps cannot be set in the past of any read timestamp that has been used. This is enforced by assertions in diagnostic builds, if applications violate this rule, data consistency can be violated.

The commits to a particular data item must be performed in timestamp order. If applications violate this rule, data consistency can be violated. Committing an update without a timestamp truncates the update's timestamp history and limits repeatable reads: no earlier version of the update will be returned regardless of the setting of the read timestamp.

The WT_SESSION::prepare_transaction API is designed to be used in conjunction with timestamps and assigns a prepare timestamp to the transaction, which will be used for visibility checks until the transaction is committed or aborted. Once a transaction has been prepared the only other operations that can be completed are WT_SESSION::commit_transaction or WT_SESSION::rollback_transaction. The WT_SESSION::prepare_transaction API only guarantees that transactional conflicts will not cause the transaction to rollback - it does not guarantee that the transactions updates are durable. If a read operation encounters an update from a prepared transaction a WT_PREPARE_CONFLICT error will be returned indicating that it is not possible to choose a version of data to return until a prepared transaction is resolved, it is reasonable to retry such operations.

Durability of the data updates performed by a prepared transaction, on tables configured with log=(enabled=false), can be controlled by specifying a durable timestamp during WT_SESSION::commit_transaction. Checkpoint will consider the durable timestamp, instead of commit timestamp for persisting the data updates. If the durable timestamp is not specified, then the commit timestamp will be considered as the durable timestamp.

There are a number of constraints around assigning timestamps for running transactions - the table below summarizes those constraints:

API Prepared Constraint Enforced Description
Prepare During prepare_timestamp >= stable_timestamp Y None
Commit No commit_timestamp > stable_timestamp Y None
Commit Yes commit_timestamp >= prepare_timestamp Y The commit timestamp may be older than the oldest timestamp at the time of commit.
Commit Yes durable_timestamp > stable_timestamp Y None
Commit Yes durable_timestamp != 0 || commit_timestamp > stable_timestamp N If no durable timestamp is given when committing a prepared transaction, the commit timestamp must be greater than the stable timestamp.

Automatic rounding of timestamps

Applications setting timestamps for a transaction have to comply with the constraints based on the global timestamp state. In order to be compliant with the constraints applications need to query the global timestamp state and check their timestamps for compliance and adjust timestamps if required. To simplify the burden on applications related to rounding up timestamps WiredTiger supports automatic rounding of timestamps in some scenarios.

Applications can configure roundup_timestamps=(prepared=true,read=true) with WT_SESSION::begin_transaction.

The configuration roundup_timestamps=(prepared=true) will be valid only for prepared transactions. It indicates that the prepare timestamp could be rounded up to the oldest timestamp, if the prepare timestamp is less than the oldest timestamp. This setting also indicates that the commit timestamp of the transaction could be rounded up to the prepare timestamp, if the commit timestamp is less than the prepare timestamp. Based on the timestamps values and constraints, enabling this configuration could result in only one of timestamps being rounded up. For example, for the timestamp values prepare_timestamp=100, commit_timestamp=300, oldest_timestamp=200 with configuration roundup_timestamps=(prepared=true) only the prepare timestamp will be rounded up to the oldest timestamp and the commit timestamp will not be adjusted and the result will be prepare_timestamp=200, commit_timestamp=300, oldest_timestamp=200. For cases where both the prepare timestamp and the commit timestamp needs to be rounded up, first the prepare timestamp will be rounded to the oldest timestamp and then the commit timestamp will be rounded up to the new prepare timestamp. For example, for the timestamp values prepare_timestamp=100, commit_timestamp=150, oldest_timestamp=200 with configuration roundup_timestamps=(prepared=true), the prepare timestamp is rounded up to the oldest timestamp, as part of the WT_SESSION::prepare_transaction, as prepare_timestamp=200 and subsequently as part of WT_SESSION::commit_transaction, the commit timestamp is rounded up to the new prepare timestamp as commit_timestamp=200.

Configuring roundup_timestamps=(prepared=true) also allows setting the prepared timestamps smaller than or equal to the latest active read timestamps. Use this feature carefully as it may break repeated read. For example, consider a transaction with a read timestamp set to 30 and a key that has a value with timestamp 20. Before a prepared transaction (in another thread), reading the key returns the value. Now a prepared transaction elsewhere in the system sets a prepared timestamp at 30. The reader, trying to do a repeat read of the key, gets return a prepared conflict error instead of the value.

Configuring roundup_timestamps=(read=true) causes the read timestamp to be rounded up to the oldest timestamp, if the read timestamp is greater than the oldest timestamp no change will be made.

Managing global timestamp state

Applications that use timestamps need to manage some global state in order to allow WiredTiger to clean up old updates, and not make new updates durable until it is safe to do so. That state is managed using the WT_CONNECTION::set_timestamp API.

Setting an oldest timestamp in WT_CONNECTION::set_timestamp indicates that future read timestamps will be at least as recent as the oldest timestamp, so WiredTiger can discard history before the specified point. It is critical that the oldest timestamp update frequently or the cache can become full of updates, reducing performance.

Setting a stable timestamp in WT_CONNECTION::set_timestamp indicates a known stable location that is sufficient for durability. During a checkpoint the state of a table will be saved only as of the stable timestamp. Newer updates after that stable timestamp will not be included in the checkpoint. That can be overridden in the call to WT_SESSION::checkpoint. It is expected that the stable timestamp is updated frequently. Setting a stable location provides the ability, if needed, to rollback to this location by placing a call to WT_CONNECTION::rollback_to_stable. With the rollback, however, WiredTiger does not automatically reset the maximum commit timestamp it is tracking. The application should explicitly do so by setting a commit timestamp in WT_CONNECTION::set_timestamp.

Timestamp Description Constraint
all_committed The oldest timestamp at which all previous write transactions have committed.
last_checkpoint The point at which the last checkpoint ran. If no checkpoint has run it's value will be 0. last_checkpoint <= stable timestamp
oldest Point in time readers cannot be created using a timestamp older than the oldest timestamp, as explained above modification history is discarded prior to the oldest timestamp. This timestamp can be set via the API. 0 <= oldest <= stable
oldest_reader The timestamp of the oldest currently active read transaction, if there is no current read transaction then querying for the oldest_reader with WT_CONNECTION::query_timestamp will return WT_NOTFOUND.
pinned Minimum of the oldest_reader and oldest timestamp.
recovery The stable timestamp used, if any, in the most recent checkpoint prior to the last shutdown.
stable Any active transaction with a commit timestamp less than or equal to the current stable timestamp will not be able to modify data, except in the instance of prepared transactions. This timestamp can be set via the API. stable >= oldest

Timestamp support in the extension API

The extension API, used by modules that extend WiredTiger via WT_CONNECTION::get_extension_api, is not timestamp-aware. In particular, WT_EXTENSION_API::transaction_oldest and WT_EXTENSION_API::transaction_visible do not take timestamps into account. Extensions relying on these functions may not work correctly with timestamps.

int rollback_transaction(WT_SESSION *session, const char *config)
Roll back the current transaction.
Conflict between concurrent operations.
int open_cursor(WT_SESSION *session, const char *uri, WT_HANDLE_NULLABLE(WT_CURSOR) *to_dup, const char *config, WT_CURSOR **cursorp)
Open a new cursor on a data source or duplicate an existing cursor.
int search(WT_CURSOR *cursor)
Return the record matching the key.
int commit_transaction(WT_SESSION *session, const char *config)
Commit the current transaction.
int open_session(WT_CONNECTION *connection, WT_EVENT_HANDLER *event_handler, const char *config, WT_SESSION **sessionp)
Open a session.
int begin_transaction(WT_SESSION *session, const char *config)
Start a transaction in this session.
void set_value(WT_CURSOR *cursor,...)
Set the value for the next operation.
int reconfigure(WT_SESSION *session, const char *config)
Reconfigure a session handle.
int update(WT_CURSOR *cursor)
Update an existing record and optionally insert a record.
void set_key(WT_CURSOR *cursor,...)
Set the key for the next operation.
int reset_snapshot(WT_SESSION *session)
Reset the snapshot.