Beacon 2.0.0 release

[robinskil]

Beacon 2.0.0 is the next major release. Version 1.8.0 came out on 2026-06-29.
This issue tracks every task before the 2.0.0 release.

Two pre-releases are done. The third pre-release is planned for this week.

Pre-releases

Tag	Date	Contents
v2.0.0-rc1	2026-08-03	nd execution spine, redb object store, wire contract move, new query editor
v2.0.0-rc2	2026-08-03	Docs for 2.0.0-rc2. Build fixes for Linux and OpenSSL.
v2.0.0-rc3	planned 2026-08-21	See #390

Three problems come from the first two pre-releases.

• CHANGELOG.md holds no section for 2.0.0-rc.2. Add one, or fold the content into the 2.0.0 section.
• The release notes of rc2 hold only a compare link. Write real notes for each later pre-release.
• The tag name and the crate version disagree. The tags read v2.0.0-rc1 and v2.0.0-rc2. Cargo.toml reads 2.0.0-rc.2. The workflow comments read v2.0.0-rc.1. Pick one form and use it.

Features in 2.0.0

Engine

• N-dimensional execution spine for Zarr and netCDF (N-dimensional execution spine: flow zarr/netcdf through beacon.nd-encoded batches #328)
• nd projection pushdown (nd projection pushdown: evaluate element-wise projections before the broadcast #329)
• nd filter pushdown (nd filter pushdown: run WHERE predicates before the broadcast #330)
• An nd scan across every partition, through a work queue (Divide an nd scan across every partition with a work queue #371)
• A single-file database, through a redb object store (Features/redb object store #332)
• A persistent statistics store per file, plus file pruning at scan time (Persistent per-file statistics: a store, a background collector, and scan-time file pruning #347)
• Column ranges from netCDF and HDF5 files (Prune HDF5 and netCDF files with their recorded statistics #360)
• Column ranges from Zarr stores (File statistics: Zarr supplies column ranges #362)
• One provider for every file scan (One provider for every file scan, pruning before it reads #364)
• A schema cache, so a query reads each file once (Keep each file's schema, so a query does not read the file twice #375)
• 123 spatial functions with PostGIS names (Use datafusion-spatial for 123 spatial functions with PostGIS names #379)

Formats

• Read Apache Iceberg tables (Read Apache Iceberg tables: external tables and read_iceberg #357)
• Read Icechunk repositories (Read Icechunk repositories through the Zarr reader #356)
• A pure-Rust netCDF reader behind a flag (feat(netcdf): add the oxcdf pure-Rust reader behind a flag #337)
• A pure-Rust HDF5 reader behind a flag (Optional pure-Rust HDF5 reader via oxcdf-hdf5 #358)
• A correct GeoParquet scan, plus bounding box pushdown (Fix the GeoParquet scan, and add bounding box pushdown #383)

Clients and interfaces

• beacon-mcp, an MCP server for LLM clients (Add MCP server (beacon-mcp): expose tables to Claude over streamable HTTP #323, MCP: export_query tool + guard rails for large results #325)
• A TypeScript client with an Arrow decoder and streams (Expose Arrow decoder and streaming capabilities in TypeScript SDK #346)
• A rebuilt admin UI query editor (Move the wire contract out of the runtime, gate the metadata schemas, and rebuild the query editor #333)
• The admin UI renders a result from the Arrow columns (Render admin UI results from Arrow columns #381)
• CREATE SECRET for object store credentials
• ATTACH for a remote Beacon catalog

Open work, wanted in 2.0.0

• Close the OIDC claim bypass, and update the dependencies (Close an OIDC claim bypass and update dependencies for security #385)
• Change a setting with SET, not only with a restart (Change a setting with SET, not only with a restart #372, Allow runtime settings to be changed with SET, not only env vars #359)
• Enforce the Rust format in CI (Enforce Rust formatting: CI gate, pre-commit hook, rustfmt.toml #327)
• Release the MCP documentation page. UNRELEASED_PAGES in docs/.vitepress/config.mts still hides it, but the feature ships.

Open work, decide before the release

• Read GeoTIFF and TIFF through the nd pipeline (Read GeoTIFF and TIFF through the nd pipeline #363)
• Async query jobs, with a poll request (Add async query jobs with incremental polling #311)
• Zarr output, gridded and flat (Zarr output: write gridded and flat Zarr v3 stores #331, draft)
• Mask functions in SELECT (Add util functions for masking in select #97)
• Dependency updates (Bump nanoid from 3.3.15 to 3.3.18 in /beacon-clients #386, Bump postcss from 8.5.6 to 8.5.26 in /docs #387, Bump js-yaml and @redocly/openapi-core in /beacon-clients #388)

Bugs to fix

Blockers

Both bugs return a wrong result, not a slow result. A major release must not ship them.

• ODV and BBF scans apply only part of a pushed-down projection (ODV and BBF scans apply only part of a pushed-down projection #382)
• A schema merge in five formats depends on the file the disk answers first (A schema merge in five formats depends on which file the disk answers first #377)

Wanted, not a blocker

• A multi-argument coalesce returns an empty result (JSON query: a multi-argument coalesce returns an empty result on the exv collection #366)
• ODV output writes an incorrect header (ODV Output Incorrect Header #131)
• The datasets store cache misses a file from a bind mount (Datasets store cache misses runtime files added via bind mount / network FS (inotify-only invalidation) #316)
• The API returns an unclear body for an empty result (API json responses for empty structs/results #75)
• netCDF output writes variable attributes as variables (Variable attributes written as variables in the netCDF #83)

Incompatible changes

State each change in the release notes, and in an upgrade page.

• One product, one name. "Beacon Data Lake" and "BeaconDB" are gone.
• beacon-datalake/ is now beacon-server/. beacon-datalake-clients/ is now beacon-clients/.
• BEACON_S3_DATA_LAKE is now BEACON_S3_DATASETS. The old name still works.
• The minimum Rust version is 1.94, up from 1.91.
• A build from source needs PROJ 9.6.2 and pkg-config for ST_Transform.
• The beacondb wheel is no longer published.

Tests

Complete tests gate the release. #399 plans the work.

• Run over 1000 Python tests against the HTTP API and BeaconDB (Release tests: Python tests for the HTTP API and BeaconDB #399)
• Cover each file format, and each file shape inside a format
• Write the what, why and how table in integration-tests/README.md

Release steps

1. Merge every blocker above.
2. Write the 2.0.0 section in CHANGELOG.md.
3. Set the version to 2.0.0 in Cargo.toml.
4. Set the version to 2.0.0 in beacon-clients/beacon-ts/package.json.
5. Set the version to 2.0.0 in beacon-clients/beacon-datalake-cli/pyproject.toml.
6. Set the version to 2.0.0 in beacon-db/beacon-db-py/pyproject.toml. The wheel stays unpublished.
7. Copy docs/docs/2.0.0-rc2/ to docs/docs/2.0.0/.
8. Update every 2.0.0-rc2 link in docs/.vitepress/config.mts.
9. Set LATEST_VERSION to 2.0.0 in docs/.vitepress/theme/version.js.
10. Run cargo +1.94 fmt --check and cargo +1.94 clippy --workspace.
11. Run cargo +1.94 test --workspace.
12. Run the integration tests with integration-tests/run.sh.
13. Run cargo audit. Record every advisory that stays open.
14. Run the full test matrix of Release tests: Python tests for the HTTP API and BeaconDB #399. Record the result here.
15. Tag v2.0.0, and push the tag.
16. Check the three publish workflows.
17. Check the container image for linux/amd64 and linux/arm64.
18. Check @beacon/client on the npm latest tag.
19. Check beacon-datalake-cli on PyPI.
20. Write the GitHub release notes.
21. Upgrade the public demo node at beacon-wod.maris.nl.
22. Update the example repository.

A v* tag starts three workflows: publish-beacon-server, publish-beacon-datalake-cli and
publish-beacon-ts. The tag is the only trigger, so a version file with a wrong value fails the run.

Out of scope

These issues move to 2.1.0.

• A registry-backed scan source (A registry-backed scan source: plan a file list without listing or opening the files #361)
• Bloom filters for equality predicates (File statistics: bloom filters for equality predicates #352)
• One statistics store, on beacon.db (Consolidate the in-memory file statistics cache onto the beacon.db statistics store #368)
• Column ranges for CSV and Arrow IPC (File statistics: CSV supplies no column ranges #349, File statistics: Arrow IPC supplies no column ranges #350)
• VirtualiZarr support (Support VirtualliZarr #112)
• Compact ODV format (Compact ODV format #193)
• An external table refresh from S3 events (S3 Events External Table Refresh #218)
• Optional GPU acceleration (Explore optional GPU acceleration via libcudf-rs #231)
• A code coverage badge (Add Code Coverage Reporting & Badge #137)
• Criterion benchmarks (Add Performance Benchmarking (Criterion) #136)
• Better error handling and logs (Improve error handling & logging across all crates #251)