DuckDB 2.0.0 LTS “Cinnamon Teal”

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

DuckDB has released version 2.0.0 LTS, a feature-focused update that adds a more mature Quack client-server protocol, asynchronous file I/O and full integration of the VARIANT data type. The project says it will support 2.0 until at least October 2027; some client libraries and extensions may become available later than the core release.

DuckDB released version 2.0.0 LTS on October 21, 2026, introducing a feature-focused update with an improved Quack client-server protocol, asynchronous I/O and expanded support for semi-structured data through the VARIANT type. The project says the release, codenamed “Cinnamon Teal,” will receive long-term support until at least October 2027, giving users a supported option as they evaluate its new capabilities and breaking changes.

The DuckDB team describes 2.0 as its biggest feature release so far, developed through close to 20,000 commits since version 1.5. That characterization comes from the project; the release announcement presents the update as focused on new features rather than the stability and backward-compatibility emphasis associated with DuckDB 1.0. The team also points users to a migration guide and full release notes because the release includes breaking changes.

On the connectivity side, 2.0 ships a more polished version of Quack, DuckDB’s native protocol for client-server use, with what the team describes as substantial performance and stability improvements. A new CONNECT statement lets users attach to and switch to a remote database in one command. DuckDB says CONNECT can also work with compatible remote databases such as PostgreSQL and MySQL; a remote pushdown optimizer can send queries to those servers rather than transferring tables to DuckDB first.

The release also adds asynchronous I/O for Parquet and CSV files and DuckDB’s own storage format. The project says the I/O layer can issue background reads and keep several requests in flight, which is intended to improve parallelism for remote data access, including on object storage. Separately, DuckDB 2.0 fully integrates the VARIANT type, which first appeared in version 1.5. It stores differently shaped, semi-structured data in a column and supports structural detection, storage and execution features described by the team as “shredding.”

New Paths for Remote DuckDB Use

The release expands DuckDB beyond workflows centered on a database embedded in a single application or process. Quack and CONNECT give users a more direct route to remote database sessions, while the stated pushdown support for PostgreSQL and MySQL could reduce data movement for compatible queries. How much that helps will depend on the database, query and workload; the announcement does not provide a general performance figure for those use cases.

Asynchronous reads matter for teams querying files remotely, where data access can hold back a query even when processing capacity is available. DuckDB says the change allows more read requests to run concurrently, but the release announcement does not establish a speedup that applies to every dataset or storage setup. Full VARIANT integration is relevant to users who need to query records with changing or inconsistent structures without treating all such data as plain text. Together, these changes broaden the kinds of data sources and workloads DuckDB says it can support.

The LTS schedule gives organizations a defined support window to plan testing and upgrades. That matters because 2.0 includes breaking changes: users may need to assess compatibility before moving production workloads rather than treating the new version as a drop-in update.

Three Supported Release Lines

DuckDB says version 2.0 is its latest long-term support release, with support scheduled through at least October 2027. The previous LTS version, 1.4, remains supported until November 17, 2026. The project has also extended support for version 1.5 until March 2027, effectively giving that line LTS status as well. The three support windows overlap, allowing users time to test the new release while continuing on an earlier version.

The announcement builds on Quack’s introduction in DuckDB 1.5.3, released in May, when the project added the protocol and an extension that lets a DuckDB process serve databases. Version 2.0 refines that functionality and adds CONNECT. The release announcement says the CLI and Python client are already available, while other client libraries, including R and Rust, and some extensions may take a few days to arrive because they require additional changes and review.

““DuckDB 2.0 is above all a feature release.””

— DuckDB team

Migration and Client Availability

The announcement identifies breaking changes but the supplied release summary does not enumerate them. Users should consult the full release notes and migration guidance before upgrading; the extent of changes needed will depend on their existing applications, queries and extensions.

Availability is not uniform across the ecosystem on announcement day. DuckDB says the CLI and Python client are already available, but R and Rust libraries and some extensions may follow over the next few days. The post also describes improvements to Quack and asynchronous I/O without giving a single benchmark result or performance guarantee covering all environments. Results on remote databases and object storage remain workload-dependent.

For security, the project says Quack listens only on localhost by default and advises putting a server behind a TLS-terminating reverse proxy when accepting connections from other machines. Its direct-exposure option, allow_other_hostname, is described as intended only for testing. The announcement does not provide deployment-specific security guidance beyond those cautions.

Testing Before the Upgrade

Users considering 2.0 can review the project’s installation instructions, release notes, migration guidance and documentation page for new features, then test the changes against their own databases and extensions. Those relying on R, Rust or affected extensions may need to wait for the relevant packages to be published and reviewed.

The project’s stated support calendar sets the next planning milestones: version 1.4 support ends on November 17, 2026, version 1.5 support continues until March 2027, and version 2.0 is supported until at least October 2027. The announcement does not give a later release date or say when each outstanding client and extension will be ready.

Key Questions

What is DuckDB 2.0.0 LTS?

It is DuckDB’s new long-term support release, announced October 21, 2026. The project says it will support the version until at least October 2027.

What are the main new features?

The release adds a more polished Quack client-server protocol, a CONNECT statement for remote database sessions, asynchronous I/O for common file formats and DuckDB storage, and fuller integration of the VARIANT type.

Does DuckDB 2.0 work with PostgreSQL and MySQL?

The project says CONNECT can point a session to compatible remote databases, including PostgreSQL and MySQL. It also says a remote pushdown optimizer can send queries to those servers. The announcement does not specify that every query or setup will be supported.

Should users upgrade immediately?

The announcement includes breaking changes, so users should check the release notes and migration guidance and test their applications before upgrading. The existing 1.4 and 1.5 support windows provide time to evaluate the new release.

Are all DuckDB 2.0 clients available now?

No. DuckDB says the CLI and Python client are available, while some extensions and the R and Rust client libraries may take a few days to arrive.

Source: hn