Apache Kudu Transaction Semantics

This is a brief introduction to Kudu’s transaction and consistency semantics. Kudu's core philosophy is to provide transactions with simple, strong semantics, without sacrificing performance or the ability to tune to different requirements. Kudu’s transactional semantics and architecture are inspired by state-of-the-art systems such as Spanner and Calvin. For an in-depth technical exposition of what is mentioned here, see the technical report.

Kudu currently allows the following operations:

  • Scans are read operations that can traverse multiple tablets and read information with some consistency or correctness guarantees. Scans can also perform time-travel reads. That is, you can set a scan timestamp from the past and get back results that reflect the state of the storage engine at that point in time.

    Write operations are sets of rows to be inserted, updated, or deleted in the storage engine, in a single tablet with multiple replicas. Write operations do not have separate "read sets", that is, they do not scan existing data before performing the write. Each write is only concerned with the previous state of the rows that are about to change. Writes are not "committed" explicitly by the user. Instead, they are committed automatically by the system, after completion.

While Kudu is designed to eventually be fully ACID (Atomic, Consistent, Isolated, Durable), multi-tablet transactions have not yet been implemented. As such, the following discussion focuses on single-tablet write operations, and only briefly touches multi-tablet reads.

Single Tablet Write Operations

Kudu employs Multiversion Concurrency Control (MVCC) and the Raft consensus algorithm. Each write operation in Kudu must go through the following order of operations:

  1. The tablet's leader acquires all locks for the rows that it will change.
  2. The leader assigns the write a timestamp before the write is submitted for replication. This timestamp will be the write’s tag in MVCC.
  3. After a majority of replicas have acknowledged the write, the rows are changed.
  4. After the changes are complete, they are made visible to concurrent writes and reads, atomically.

All replicas of a tablet observe the same process. Therefore, if a write operation is assigned timestamp n, and changes row x, a second write operation at timestamp m > n is guaranteed to see the new value of x.

This strict ordering of lock acquisition and timestamp assignment is enforced to be consistent across all replicas of a tablet through consensus. Therefore, write operations are ordered with regard to clock-assigned timestamps, relative to other writes in the same tablet. In other words, writes have strict-serializable semantics.

In case of multi-row write operations, while they are Isolated and Durable in an ACID sense, they are not yet fully Atomic. The failure of a single write in a batch operation will not roll back the entire operation, but produce per-row errors.

Writing to Multiple Tablets

Kudu does not support transactions that span multiple tablets. However, consistent snapshot reads are possible (with caveats, as explained below). Writes from a Kudu client are optionally buffered in memory until they are flushed and sent to the tablet server. When a client’s session is flushed, the rows for each tablet are batched together, and sent to the tablet server which hosts the leader replica of the tablet. Since there are no inter-tablet transactions, each of these batches represents a single, independent write operation with its own timestamp. However, the client API provides the option to impose some constraints on the assigned timestamps and on how writes to different tablets are observed by clients.

Kudu was designed to be externally consistent, that is, preserving consistency when operations span multiple tablets and even multiple data centers. In practice this means that if a write operation changes item x at tablet A, and a following write operation changes item y at tablet B, you might want to enforce that if the change to y is observed, the change to x must also be observed. There are many examples where this can be important. For example, if Kudu is storing clickstreams for further analysis, and two clicks follow each other but are stored in different tablets, subsequent clicks should be assigned subsequent timestamps so that the causal relationship between them is captured.


    Kudu’s default external consistency mode is called CLIENT_PROPAGATED. This mode causes writes from a single client to be automatically externally consistent. In the clickstream scenario above, if the two clicks are submitted by different client instances, the application must manually propagate timestamps from one client to the other for the causal relationship to be captured. Timestamps between clients a and b can be propagated as follows:

    Java Client

    Call AsyncKuduClient#getLastPropagatedTimestamp() on client a, propagate the timestamp to client b, and call AsyncKuduClient#setLastPropagatedTimestamp() on client b.

    C++ Client

    Call KuduClient::GetLatestObservedTimestamp() on client a, propagate the timestamp to client b, and call KuduClient::SetLatestObservedTimestamp() on client b.

  • COMMIT_WAIT Consistency

    Kudu also has an experimental implementation of an external consistency model (used in Google’s Spanner), called COMMIT_WAIT. COMMIT_WAIT works by tightly synchronizing the clocks on all machines in the cluster. Then, when a write occurs, timestamps are assigned and the results of the write are not made visible until enough time has passed so that no other machine in the cluster could possibly assign a lower timestamp to a following write.

    When using this mode, the latency of writes is tightly tied to the accuracy of clocks on all the cluster hosts, and using this mode with loose clock synchronization causes writes to either take a long time to complete, or even time out.

    The COMMIT_WAIT consistency mode may be selected as follows:

    Java Client

    Call KuduSession#setExternalConsistencyMode(ExternalConsistencyMode.COMMIT_WAIT)

    C++ Client

    Call KuduSession::SetExternalConsistencyMode(COMMIT_WAIT)

Read Operations (Scans)

Scans are read operations performed by clients that may span one or more rows across one or more tablets. When a server receives a scan request, it takes a snapshot of the MVCC state and then proceeds in one of two ways depending on the read mode selected by the user. The mode may be selected as follows:

Java Client

Call KuduScannerBuilder#ReadMode(…​)

C++ Client

Call KuduScanner::SetReadMode()

The following modes are available in both clients:


This is the default read mode. The server takes a snapshot of the MVCC state and proceeds with the read immediately. Reads in this mode only yield 'Read Committed' isolation.


In this read mode, scans are consistent and repeatable. A timestamp for the snapshot is selected either by the server, or set explicitly by the user through KuduScanner::SetSnapshotMicros(). Explicitly setting the timestamp is recommended.

The server waits until this timestamp is 'safe'; that is, until all write operations that have a lower timestamp have completed and are visible). This delay, coupled with an external consistency method, will eventually allow Kudu to have full strict-serializable semantics for reads and writes. However, this is still a work in progress and some anomalies are still possible. Only scans in this mode can be fault-tolerant.

Selecting between read modes requires balancing the trade-offs and making a choice that fits your workload. For instance, a reporting application that needs to scan the entire database might need to perform careful accounting operations, so that scan may need to be fault-tolerant, but probably doesn’t require a to-the-microsecond up-to-date view of the database. In that case, you might choose READ_AT_SNAPSHOT and select a timestamp that is a few seconds in the past when the scan starts. On the other hand, a machine learning workload that is not ingesting the whole data set and is already statistical in nature might not require the scan to be repeatable, so you might choose READ_LATEST instead for better scan performance.

Known Issues and Limitations

There are several gaps and corner cases that currently prevent Kudu from being strictly-serializable in certain situations.


Support for COMMIT_WAIT is experimental and requires careful tuning of the time-synchronization protocol, such as NTP (Network Time Protocol). Its use in production environments is discouraged.

If external consistency is a requirement and you decide to use COMMIT_WAIT, the time-synchronization protocol needs to be tuned carefully. Each transaction will wait 2x the maximum clock error at the time of execution, which is usually in the 100 msec. to 1 sec. range with the default settings, maybe more. Thus, transactions would take at least 200 msec. to 2 sec. to complete when using the default settings and may even time out.
  • A local server should be used as a time server. We’ve performed experiments using the default NTP time source available in a Google Compute Engine data center and were able to obtain a reasonable tight max error bound, usually varying between 12-17 milliseconds.

  • The following parameters should be adjusted in /etc/ntp.conf to tighten the maximum error:

    • server my_server.org iburst minpoll 1 maxpoll 8

    • tinker dispersion 500

    • tinker allan 0

Reads (Scans)

  • On a leader change, READ_AT_SNAPSHOT scans at a snapshot whose timestamp is beyond the last write, may yield non-repeatable reads (see KUDU-1188).

    If repeatable snapshot reads are a requirement, use READ_AT_SNAPSHOT with a timestamp that is slightly in the past (between 2-5 seconds, ideally). This will circumvent the anomaly described above. Even when the anomaly has been addressed, back-dating the timestamp will always make scans faster, since they are unlikely to block.
  • Impala scans are currently performed as READ_LATEST and have no consistency guarantees.

  • In AUTO_BACKGROUND_FLUSH mode, or when using "async" flushing mechanisms, writes applied to a single client session may get reordered due to the concurrency of flushing the data to the server. This is particularly noticeable if a single row is quickly updated with different values in succession. This phenomenon affects all client API implementations. Workarounds are described in the respective API documentation for FlushMode or AsyncKuduSession. See KUDU-1767.