use GCS mirror for Maven, fix benchmarks version, other small changes#166
Conversation
| -Daether.connector.basic.downstreamThreads=1 | ||
| -Daether.transport.http.retryHandler.count=5 | ||
| -Daether.transport.http.retryHandler.interval=10000 | ||
| -Dmaven.wagon.http.retryHandler.count=5 |
There was a problem hiding this comment.
See NVIDIA/cuvs#2253 (comment) for an explanation of these flags.
I believe comments are not allowed in this file because its contents are passed literally to mvn.
jakirkham
left a comment
There was a problem hiding this comment.
Thanks James! 🙏
Had a couple questions below. Most can be handled separately
Would just double check the copyright one
| @@ -1,4 +1,4 @@ | |||
| # SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. | |||
| # SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. | |||
There was a problem hiding this comment.
What is the context behind this addition?
There was a problem hiding this comment.
It was added automatically by the verify-codeowners pre-commit hook.
This became the default in February (rapidsai/pre-commit-hooks#120) based on internal conversations about the preferred form.
It's showing up here now because:
- that change went into
rapidsai/pre-commit-hooksv1.5.0 and this repo was using an older version until this PR - the hook only changes copyright statements on files as they're touched, not all at once, and this PR touches that file
If you have questions about the history behind rapidsai/pre-commit-hooks#120 you can talk to @KyleFromNVIDIA . But this has been the standard across RAPIDS for a few months now.
There was a problem hiding this comment.
Do we need this in other files then?
A few files here changed, but not all have this particular wording
There was a problem hiding this comment.
I don't know why this project doesn't include .xml files in its configuration for that hook. It's probably unintentional, I'd be happy to approve a PR adding them if you want to put one up.
| <groupId>com.nvidia.cuvs.lucene.benchmarks</groupId> | ||
| <artifactId>cuvs-lucene-benchmarks</artifactId> | ||
| <!--CUVS_JAVA#VERSION_UPDATE_MARKER_START--><version>26.02.0</version><!--CUVS_JAVA#VERSION_UPDATE_MARKER_END--> | ||
| <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END--> |
There was a problem hiding this comment.
Is there anywhere else that needs this change?
There was a problem hiding this comment.
Possibly needs changes in ci/release/update-version.sh?
There was a problem hiding this comment.
is there anywhere else that needs this change?
No.
$ git grep -E '26\.|25\.'
.github/workflows/build.yaml: container_image: "rapidsai/ci-conda:26.08-cuda${{ matrix.cuda_version }}-ubuntu24.04-py3.13"
.github/workflows/pr.yaml: container_image: "rapidsai/ci-conda:26.08-cuda${{ matrix.cuda_version }}-ubuntu24.04-py3.13"
.github/workflows/test.yaml: container_image: "rapidsai/ci-conda:26.08-cuda${{ matrix.cuda_version }}-ubuntu24.04-py3.13"
.pre-commit-config.yaml: rev: v1.26.1
VERSION:26.08.00
bench/pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
bench/pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
build.sh:VERSION="26.08.0" # Note: The version is updated automatically when ci/release/update-version.sh is invoked
ci/release/update-version.sh:# Strip leading 0s in versions, so e.g. '25.10.00' becomes '25.10.0'
conda/environments/all_cuda-129_arch-aarch64.yaml:- libcuvs==26.08.*
conda/environments/all_cuda-129_arch-x86_64.yaml:- libcuvs==26.08.*
conda/environments/all_cuda-132_arch-aarch64.yaml:- libcuvs==26.08.*
conda/environments/all_cuda-132_arch-x86_64.yaml:- libcuvs==26.08.*
dependencies.yaml: - libcuvs==26.08.*
examples/README.md:docker run --rm --gpus all --pull=always --volume $PWD:$PWD --workdir $PWD -it rapidsai/ci-conda:26.08-cuda13.2.0-ubuntu24.04-py3.13
examples/README.md:mvn clean install && java -Djava.util.logging.config.file=src/main/resources/logging.properties -cp target/examples-26.08.0-jar-with-merged-services.jar com.nvidia.cuvs.lucene.examples.AcceleratedHnswExample
examples/README.md:mvn clean install && java -Djava.util.logging.config.file=src/main/resources/logging.properties -cp target/examples-26.08.0-jar-with-merged-services.jar com.nvidia.cuvs.lucene.examples.IndexAndSearchonGPUExample
examples/pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
examples/pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
pom.xml: <!--CUVS_LUCENE#VERSION_UPDATE_MARKER_START--><version>26.08.0</version><!--CUVS_LUCENE#VERSION_UPDATE_MARKER_END-->
src/main/java/com/nvidia/cuvs/lucene/CuVS2510GPUSearchCodec.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/CuVS2510GPUVectorsFormat.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/CuVS2510GPUVectorsReader.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/CuVS2510GPUVectorsWriter.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/FilterCuVSServiceProvider.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/GPUBuiltHnswGraph.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/GPUFieldWriter.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/GPUIndex.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/GPUKnnFloatVectorQuery.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/GPUPerLeafCuVSKnnCollector.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/IndexInputInputStream.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/IndexOutputOutputStream.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/Lucene101AcceleratedHNSWCodec.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/Lucene99AcceleratedHNSWVectorsFormat.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/Lucene99AcceleratedHNSWVectorsWriter.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWBinaryQuantizedCodec.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWBinaryQuantizedVectorsFormat.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWBinaryQuantizedVectorsWriter.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWScalarQuantizedCodec.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWScalarQuantizedVectorsFormat.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneAcceleratedHNSWScalarQuantizedVectorsWriter.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/LuceneProvider.java: * @since 25.12
src/main/java/com/nvidia/cuvs/lucene/ThreadLocalCuVSResourcesProvider.java: * @since 26.02
src/main/java/com/nvidia/cuvs/lucene/Utils.java: * @since 25.10
src/main/java/com/nvidia/cuvs/lucene/package-info.java: * @since 25.10
src/test/java/com/nvidia/cuvs/lucene/TestBackCompat.java: * @since 25.12Possibly needs changes in ci/release/update-version.sh?
I don't think so. The file was covered already, the pattern was just wrong.
bash ./ci/release/update-version.sh '26.10.00' --run-context=main
git grep -E '25\.|26\.'
# (same result as above, but with all 26.08 changed to 26.10)|
/merge |
The merge of origin/main (c840848) combined two independently-added <repositories> blocks: the GCS-mirror + Maven Central block from NVIDIA#166 (releases only, snapshots disabled) and the central-snapshots block from the snapshot-build change (snapshots enabled). Two <repositories> tags is invalid XML, so the POM failed to parse. Consolidate both into a single <repositories> block, preserving all three repositories: the GCS mirror and Maven Central for releases, and central-snapshots for snapshot resolution.
Fixes #145
Contributes to rapidsai/build-planning#297
The root cause of #145 appears to be that we were getting rate-limited by Maven Central. Similar to NVIDIA/cuvs#2253, this proposes fixing that by using the same read-only Maven mirror that Apache Orc, Lucene, Spark and others use in their builds (https://storage-download.googleapis.com/maven-central/index.html).
Other changes:
mvnoptions to request packages more slowly and wait longer between retriespre-commithooks w/pre-commit autoupdatecheck-xmlhook to validate thatpom.xmlare valid XML docs