ContextQMD
Back to Clickhouse

2024 Changelog

docs/changelogs/v24.8.9.95-lts.md

26.4.1.1-new12.6 KB
Original Source

2024 Changelog

ClickHouse release v24.8.9.95-lts (386a2469df1) FIXME as compared to v24.8.8.17-lts (81036bd118b)

Backward Incompatible Change

Performance Improvement

Improvement

  • Backported in #72568: Fix the metadata_version record in ZooKeeper in restarting thread rather than in attach thread. #70297 (Miсhael Stetsyuk).
  • Backported in #73561: Fix the issue where ClickHouse in Docker containers printed "get_mempolicy: Operation not permitted" into stderr due to restricted syscalls. #70900 (filimonov).
  • Backported in #72671: Better memory usage approximation of cached marks to reduce total memory usage of the cache. #72630 (Antonio Andelic).

Bug Fix (user-visible misbehavior in an official stable release)

  • Backported in #72410: The parts deduplicated during ATTACH PART query don't get stuck with the attaching_ prefix anymore. #65636 (Kirill).
  • Backported in #73503: Fix #68239 SAMPLE n where n is an integer. #68499 (Denis Hananein).
  • Backported in #73356: Fixed zombie processes after after library bridge crash. #71301 (MikhailBurdukov).
  • Backported in #73355: Fix error column "attgenerated" does not exist for old PostgreSQL, fix #60651. #71396 (0xMihalich).
  • Backported in #72726: Fix NoSuchKey error during transaction rollback when creating a directory fails for the palin_rewritable disk. #71439 (Julia Kartseva).
  • Backported in #72170: Fix counting column size in wide part for Dynamic and JSON types. #71526 (Pavel Kruglov).
  • Backported in #72269: Fix serialization of Dynamic values in Pretty JSON formats. #71923 (Pavel Kruglov).
  • Backported in #72181: Fix rows_processed column in system.s3/azure_queue_log broken in 24.6. Closes #69975. #71946 (Kseniia Sumarokova).
  • Backported in #73397: Add inferred format name to create query in File/S3/URL/HDFS/Azure engines. Previously the format name was inferred each time the server was restarted, and if the specified data files were removed, it led to errors during server startup. #72108 (Pavel Kruglov).
  • Backported in #72305: Fix bugs when using UDF in join on expression with the old analyzer. #72179 (Raúl Marín).
  • Backported in #72250: Fix a bug where min_age_to_force_merge_on_partition_only was getting stuck trying to merge down the same partition repeatedly that was already merged to a single part and not merging partitions that had multiple parts. #72209 (Christoph Wurm).
  • Backported in #72323: Fixed data race in GraceHashJoin as the result of which some rows might be missing in the join output. #72233 (Nikita Taranov).
  • Backported in #72312: Fixed data race when ColumnDynamic::dumpStructure() is called concurrently e.g. in ConcurrentHashJoin constructor. #72278 (Nikita Taranov).
  • Backported in #72523: Fixed mismatched types in several cases after applying optimize_functions_to_subcolumns. #72394 (Anton Popov).
  • Backported in #72502: Fix failure on parsing BACKUP DATABASE db EXCEPT TABLES db.table queries. #72429 (Konstantin Bogdanov).
  • Backported in #72952: Don't allow creating empty Variant. #72454 (Pavel Kruglov).
  • Backported in #72709: Fix parsing a glob with one element. #72572 (Konstantin Bogdanov).
  • Backported in #72926: Fix query generation for the follower server in case of a distributed query with ARRAY JOIN. Fixes #69276. #72608 (Dmitry Novik).
  • Backported in #72846: Fixed inconsistent metadata when adding a new replica to a Replicated database that has a table created with flatten_nested=0. #72685 (Alexander Tokmakov).
  • Backported in #73314: Fix the situation when a user can't run REVOKE ALL ON *.* because of implicit grants in the target access entity. #72872 (pufit).
  • Backported in #72979: Fix potential crash during backup restore. #72947 (Kseniia Sumarokova).
  • Backported in #73155: Fix crash in transactions with unsupported storage. #73045 (Raúl Marín).
  • Backported in #73124: Fix possible overestimate memory tracking (difference between MemoryTracking and MemoryResident kept growing). #73081 (Azat Khuzhin).
  • Backported in #73106: Check for duplicate JSON keys during Tuple parsing. Previously it could lead to logical error Invalid number of rows in Chunk during parsing. #73082 (Pavel Kruglov).
  • Backported in #73610: Fix ObjectStorageQueue with ZooKeeper and older Keeper. #73420 (Antonio Andelic).
  • Backported in #73620: Fixed error in function tupleElement which may appear in some cases for tuples with LowCardinality elelments and enabled setting optimize_functions_to_subcolumns. #73548 (Anton Popov).
  • Backported in #73759: Fix parsing enum glob followed by range one. Fixes #73473. #73569 (Konstantin Bogdanov).
  • Backported in #73684: Do not interpret queries in EXPLAIN SYNTAX to avoid logical errors with incorrect processing stage for distributed queries. Fixes #65205. #73634 (Dmitry Novik).
  • Backported in #73730: Fix crash in LIMIT BY COLUMNS. #73686 (Raúl Marín).

NOT FOR CHANGELOG / INSIGNIFICANT