Skip to content

Releases: ozontech/seq-db

v0.75.0

Choose a tag to compare

@cheb0 cheb0 released this 30 Jul 07:52
8cf7b08

What's Changed

  • fix: add write permission to WAL file while copying by @cheb0 in #466
  • fix: log search error if error code is present by @cheb0 in #468
  • fix: stop sw on context cancel (search err) by @cheb0 in #467
  • perf: optimize TID entry lookup by @cheb0 in #454
  • fix: prevent unique stack traces explosion in mem profile by @cheb0 in #464
  • chore: delete IsLastLID by @cheb0 in #469
  • feat: add token block size config by @cheb0 in #439
  • perf: implement letter based skip index for token blocks by @cheb0 in #461
  • chore: remove compatibility with .meta file by @dkharms in #462
  • feat: add token frequencies by @cheb0 in #438
  • fix: active frac hanging if no disk space is left for some time by @cheb0 in #475
  • refactor: add ReaderProvider to fetch registry from cache once per re… by @forshev in #447
  • feat(async search): add export async search docs method by @forshev in #473

Full Changelog: v0.74.0...v0.75.0

v0.73.3

Choose a tag to compare

@eguguchkin eguguchkin released this 17 Jul 13:21

Full Changelog: v0.73.2...v0.73.3

  • fix(downsample): early exit from the loop (#460)

v0.73.2

Choose a tag to compare

@eguguchkin eguguchkin released this 14 Jul 10:33

Full Changelog: v0.73.1...v0.73.2

  • fix: correctly release buckets (#456)

v0.74.0

Choose a tag to compare

@dkharms dkharms released this 13 Jul 14:26
42fb1c0

What's Changed

  • feat: new api for compaction by @dkharms in #429
  • refactor: make index building reusable by @dkharms in #401
  • fix: bring back default capacity for LID block by @dkharms in #448
  • refactor: do not store ids count in offsets by @dkharms in #410
  • fix: trivial test fix by @eguguchkin in #449
  • feat: implement merge source for compaction by @dkharms in #413
  • feat: implement STCS compaction by @dkharms in #430
  • fix(downsample): early exit from the loop by @eguguchkin in #460
  • fix: add .compaction-plan to handle in-mid compaction crash by @dkharms in #453
  • fix: correctly release buckets by @dkharms in #456
  • fix: correctly handle non-existing ids by @dkharms in #458
  • fix: data loss in asyncsearch merge caused by concurrent retention deletion of fracs by @eguguchkin in #459
  • test: relax tolerances and remove flaky index-on-disk assertions by @eguguchkin in #455
  • feat: add logging for fraction lifecycle by @eguguchkin in #450

Full Changelog: v0.73.0...v0.74.0

v0.73.1

Choose a tag to compare

@eguguchkin eguguchkin released this 13 Jul 14:35

Full Changelog: v0.73.0...v0.73.1

  • fix: data loss in asyncsearch merge caused by concurrent retention deletion of fracs (#459)

v0.73.0

Choose a tag to compare

@eguguchkin eguguchkin released this 25 Jun 16:31
0b00c80

What's Changed

  • feat(fracmanager): implement fraction snapshots with wait group reference counting by @eguguchkin in #277
  • refactor: move WAL corruption metric to periodic stats worker counting broken .docs files by @eguguchkin in #433
  • perf: improve compression throughput and reduce allocations by @eguguchkin in #408
  • fix: add overlap check for TID block in pattern search by @cheb0 in #437
  • fix: add search params validation that from <= to by @cheb0 in #442
  • feat: add random search downsampling by @eguguchkin in #443

Full Changelog: v0.72.0...v0.73.0

v0.72.1

Choose a tag to compare

@cheb0 cheb0 released this 11 Jun 09:08

What's Changed

  • fix: add overlap check for TID block in pattern search by @cheb0 in #437

Full Changelog: v0.72.0...v0.72.1

v0.72.0

Choose a tag to compare

@eguguchkin eguguchkin released this 09 Jun 08:42
985c744

What's Changed

  • chore: don't evaluate skip masks in fetch after search by @forshev in #412
  • perf(skip masks): skip masks as roaring bitmaps in fetch by @forshev in #421
  • chore(skipmaskmanager): remove skip mask files at startup if no frac by @forshev in #427
  • chore(skip mask manager): use roaring v2 by @forshev in #434
  • feat: batching execution (single iterator queries) by @cheb0 in #390
  • perf: implement batch processing in iterateEvalTree by @cheb0 in #406
  • perf: implement bitpack encoding for LID and MID blocks by @cheb0 in #328
  • fix: fix corruption count not passed while replaying WAL by @cheb0 in #432
  • feat: implement LID block size config by @cheb0 in #405
  • perf: improve wildcard query perf with predicate and contains-check pushdown by @cheb0 in #397

Full Changelog: v0.71.0...v0.72.0

v0.71.0

Choose a tag to compare

@eguguchkin eguguchkin released this 29 May 13:14
003f3cc

What's Changed

  • fix: race condition fractionRegistry stats by @eguguchkin in #388
  • chore: сleanup metadata version 1 by @eguguchkin in #396
  • refactor: split .index file into multiple by @dkharms in #389
  • chore: copy corrupted WAL and docs to .broken dir by @cheb0 in #395
  • perf: find TokenTable index once by @dkharms in #425
  • feat(async-search): fetch fractions one by one to avoid global lock by @eguguchkin in #424
  • chore: entry count statistics for cache by @cheb0 in #428

Full Changelog: v0.70.0...v0.71.0

v0.70.0

Choose a tag to compare

@forshev forshev released this 10 Apr 14:26
124d460

What's Changed

Full Changelog: v0.69.0...v0.70.0