Gabriel39 opened a new pull request, #4181: URL: https://github.com/apache/doris-website/pull/4181
The Lance Catalog page lists cache settings but does not explain how to size the index cache for a query workload. Add matching English and Chinese guidance for the six supported vector index types, including row IDs, quantization, and HNSW graph overhead. The examples derive a per-BE working set from distinct cached IVF partitions and show how to choose an initial BE cache budget. The guidance also covers scalar/full-text index sharing, metadata-cache tuning, cache metrics, query working memory, and the separate disk data cache. It links to the upstream Lance performance and index-format documentation. Validation: - Both complete MDX pages compile with the repository's Docusaurus 3.6.3 MDX compiler. - Changed-file frontmatter and Markdown structure checks report no findings. - Example arithmetic and byte-valued configuration settings verified; both upstream reference URLs return HTTP 200. - Full diff reviewed and `git diff --check` passes. A full website build was not run. Scope and synchronization: - English and Chinese 4.x pages are updated together. The i18n checker reports that the existing English Lance Catalog page has no `current` counterpart; this PR intentionally extends the existing 4.x documentation rather than introducing a new versioned page. - Existing headings, URLs, frontmatter, navigation, and UI are preserved. The new sizing section has the same explicit anchor in both languages. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
