Skip to content

Latest commit

 

History

History
410 lines (289 loc) · 49.6 KB

release-8.2.0.md

File metadata and controls

410 lines (289 loc) · 49.6 KB
Raw
title summary
TiDB 8.2.0 Release Notes
Learn about the new features, compatibility changes, improvements, and bug fixes in TiDB 8.2.0.

TiDB 8.2.0 Release Notes

Release date: July 11, 2024

TiDB version: 8.2.0

Quick access: Quick start

8.2.0 introduces the following key features and improvements:

Category Feature/Enhancement Description
Reliability and Availability TiProxy supports multiple load balancing policies In TiDB v8.2.0, TiProxy evaluates and ranks TiDB nodes based on various dimensions, such as status, connection counts, health, memory, CPU, and location. According to the load balancing policy specified in the policy configuration item, TiProxy dynamically selects the optimal TiDB node to execute database operations. This optimizes overall resource usage, improves cluster performance, and increases throughput.
The parallel HashAgg algorithm of TiDB supports disk spill (GA) HashAgg is a widely used aggregation operator in TiDB for efficiently aggregating rows with the same field values. TiDB v8.0.0 introduces parallel HashAgg as an experimental feature to further enhance processing speed. When memory resources are insufficient, parallel HashAgg spills temporary sorted data to disk, avoiding potential OOM risks caused by excessive memory usage. This improves query performance while maintaining node stability. In v8.2.0, this feature becomes generally available (GA) and is enabled by default, enabling you to safely configure the concurrency of parallel HashAgg using tidb_executor_concurrency.
Improve statistics loading efficiency by up to 10 times For clusters with a large number of tables and partitions, such as SaaS or PaaS services, improvement in statistics loading efficiency can solve the problem of slow startup of TiDB instances, and increase the success rate of dynamic loading of statistics. This improvement reduces performance rollbacks caused by statistics loading failures and improves cluster stability.
DB Operations and Observability Introduce privilege control of switching resource groups As resource control is widely used, the privilege control of switching resource groups can prevent database users from abusing resources, strengthen administrators' protection of overall resource usage, and improve cluster stability.

Feature details

Performance

  • Support pushing down the following JSON functions to TiKV #50601 @dbsid

    • JSON_ARRAY_APPEND()
    • JSON_MERGE_PATCH()
    • JSON_REPLACE()

    For more information, see documentation.

  • TiDB supports parallel sorting #49217 #50746 @xzhangxian1008

    Before v8.2.0, TiDB only executes Sort operators sequentially, affecting query performance when sorting large amounts of data.

    Starting from v8.2.0, TiDB supports parallel sorting, which significantly improves sorting performance. This feature does not need manual configuration. TiDB automatically determines whether to use parallel sorting based on the value of the tidb_executor_concurrency system variable.

    For more information, see documentation.

  • The parallel HashAgg algorithm of TiDB supports disk spill (GA) #35637 @xzhangxian1008

    TiDB v8.0.0 introduces the parallel HashAgg algorithm with disk spill support as an experimental feature. In v8.2.0, this feature becomes generally available (GA). When using the parallel HashAgg algorithm, TiDB automatically triggers data spill based on memory usage, thus balancing query performance and data throughput. This feature is enabled by default. The system variable tidb_enable_parallel_hashagg_spill, which controls this feature, will be deprecated in a future release.

    For more information, see documentation.

Reliability

  • Improve statistics loading efficiency by up to 10 times #52831 @hawkingrei

    SaaS or PaaS applications can have a large number of data tables, which not only slow down the loading speed of the initial statistics, but also increase the failure rate of load synchronization under high loads. The startup time of TiDB and the accuracy of the execution plan can be affected. In v8.2.0, TiDB optimizes the process of loading statistics from multiple perspectives, such as the concurrency model and memory allocation, to reduce latency, improve throughput, and avoid slow loading of statistics that affect business scaling.

    Adaptive concurrent loading is now supported. By default, the configuration item stats-load-concurrency is set to 0, and the concurrency of statistics loading is automatically selected based on the hardware specification.

    For more information, see documentation.

Availability

  • TiProxy supports multiple load balancing policies #465 @djshow832 @xhebox

    TiProxy is the official proxy component of TiDB, located between the client and TiDB server. It provides load balancing and connection persistence functions for TiDB. Before v8.2.0, TiProxy defaults to v1.0.0, which only supports status-based and connection count-based load balancing policies for TiDB servers.

    Starting from v8.2.0, TiProxy defaults to v1.1.0 and introduces multiple load balancing policies. In addition to status-based and connection count-based policies, TiProxy supports dynamic load balancing based on health, memory, CPU, and location, improving the stability of the TiDB cluster.

    You can configure the combination and priority of load balancing policies through the policy configuration item.

    • resource: the resource priority policy performs load balancing based on the following priority order: status, health, memory, CPU, location, and connection count.
    • location: the location priority policy performs load balancing based on the following priority order: status, location, health, memory, CPU, and connection count.
    • connection: the minimum connection count priority policy performs load balancing based on the following priority order: status and connection count.

    For more information, see documentation.

SQL

  • TiDB supports the JSON schema validation function #52779 @dveeden

    Before v8.2.0, you need to rely on external tools or customized validation logic for JSON data validation, which increases the complexity of development and maintenance, and reduces development efficiency. Starting from v8.2.0, the JSON_SCHEMA_VALID() function is introduced. Using JSON_SCHEMA_VALID() in the CHECK constraint can help prevent non-conforming data from being inserted, rather than checking the data after it has been added. This function lets you verify the validity of JSON data directly in TiDB, improving the integrity and consistency of the data, and increasing the development efficiency.

    For more information, see documentation.

DB operations

  • TiUP supports deploying PD microservices #5766 @rleungx

    Starting from v8.0.0, PD supports the microservice mode. This mode splits the timestamp allocation and cluster scheduling functions of PD into separate microservices that can be deployed independently, thereby improving resource control and isolation, and reducing the impact between different services. Before v8.2.0, PD microservices can only be deployed using TiDB Operator.

    Starting from v8.2.0, PD microservices can also be deployed using TiUP. You can deploy the tso microservice and the scheduling microservice separately in a cluster to enhance PD performance scalability and address PD performance bottlenecks in large-scale clusters. It is recommended to use this mode when PD becomes a significant performance bottleneck that cannot be resolved by scaling up.

    For more information, see user documentation.

  • Add privilege control of switching resource groups #53440 @glorv

    TiDB lets users switch to other resource groups using the SET RESOURCE GROUP command or the RESOURCE_GROUP() hint, which might lead to resource group abuse by some database users. TiDB v8.2.0 introduces privilege control of switching resource groups. Only database users granted the RESOURCE_GROUP_ADMIN or RESOURCE_GROUP_USER dynamic privilege can switch to other resource groups, enhancing the protection of system resources.

    To maintain compatibility, the original behavior is retained when upgrading from earlier versions to v8.2.0 or later versions. To enable the enhanced privilege control, set the new variable tidb_resource_control_strict_mode to ON.

    For more information, see user documentation.

Observability

  • Record the reason why an execution plan is not cached #50618 @qw4990

    In some scenarios, you might want to cache most execution plans to save execution overhead and reduce latency. Currently, execution plan caching has some limitations on SQL. Execution plans of some SQL statements cannot be cached. It is difficult to identify the SQL statements that cannot be cached and the corresponding reasons.

    Therefore, starting from v8.2.0, new columns PLAN_CACHE_UNQUALIFIED and PLAN_CACHE_UNQUALIFIED_LAST_REASON are added to the system table STATEMENTS_SUMMARY to explain the reason why an execution plan cannot be cached, which can help you tune performance.

    For more information, see documentation.

Security

  • Enhance TiFlash log desensitization #8977 @JaySon-Huang

    TiDB v8.0.0 enhances the log desensitization feature, enabling you to control whether user data in TiDB logs is wrapped in markers ‹ ›. Based on the marked logs, you can decide whether to redact the marked information when displaying logs, thereby increasing the flexibility of log desensitization. In v8.2.0, TiFlash introduces a similar enhancement for log desensitization. To use this feature, set the TiFlash configuration item security.redact_info_log to marker.

    For more information, see documentation.

Data migration

  • Align TiCDC Syncpoints across multiple changefeeds #11212 @hongyunyan

    Before v8.2.0, aligning TiCDC Syncpoints across multiple changefeeds was challenging. The startTs of the changefeed had to be carefully selected when the changefeed was created, so it would align with the Syncpoints of other changefeeds. Starting from v8.2.0, Syncpoints for a changefeed are created as a multiple of the changefeed's sync-point-interval configuration. This change lets you align Syncpoints across multiple changefeeds that have the same sync-point-interval configuration, simplifying and improving the ability to align multiple downstream clusters.

    For more information, see documentation.

  • TiCDC Pulsar Sink supports using the pulsar+http and pulsar+https connection protocols #11336 @SandeepPadhi

    Before v8.2.0, TiCDC Pulsar Sink only supports pulsar and pulsar+ssl connection protocols. Starting from v8.2.0, TiCDC Pulsar Sink also supports pulsar+http and pulsar+https protocols for connections. This enhancement improves the flexibility of connecting to Pulsar.

    For more information, see documentation.

Compatibility changes

Note:

This section provides compatibility changes you need to know when you upgrade from v8.1.0 to the current version (v8.2.0). If you are upgrading from v8.0.0 or earlier versions to the current version, you might also need to check the compatibility changes introduced in intermediate versions.

Behavior changes

  • When using TiDB Lightning to import a CSV file, if you set strict-format = true to split a large CSV file into multiple small CSV files to improve concurrency and import performance, you need to explicitly specify terminator. The values can be \r, \n or \r\n. Failure to specify a line terminator might result in an exception when parsing the CSV file data. #37338 @lance6716

  • When using IMPORT INTO to import a CSV file, if you specify the SPLIT_FILE parameter to split a large CSV file into multiple small CSV files to improve concurrency and import performance, you need to explicitly specify the line terminator LINES_TERMINATED_BY. The values can be \r, \n or \r\n. Failure to specify a line terminator might result in an exception when parsing the CSV file data. #37338 @lance6716

  • Before BR v8.2.0, performing BR data restore on a cluster with TiCDC replication tasks is not supported. Starting from v8.2.0, BR relaxes the restrictions on data restoration for TiCDC: if the BackupTS (the backup time) of the data to be restored is earlier than the changefeed CheckpointTS (the timestamp that indicates the current replication progress), BR can proceed with the data restore normally. Considering that BackupTS is usually much earlier, it can be assumed that in most scenarios, BR supports restoring data for a cluster with TiCDC replication tasks. #53131 @YuJuncen

MySQL compatibility

  • Before v8.2.0, executing the CREATE USER statement with the PASSWORD REQUIRE CURRENT DEFAULT option returns an error because this option is not supported and cannot be parsed. Starting from v8.2.0, TiDB supports parsing and ignoring this option for compatibility with MySQL. #53305 @dveeden

System variables

Variable name Change type Description
tidb_analyze_distsql_scan_concurrency Modified Changes the minimum value from 1 to 0. When you set it to 0, TiDB adaptively adjusts the concurrency of the scan operation when executing the ANALYZE operation based on the cluster size.
tidb_analyze_skip_column_types Modified Starting from v8.2.0, TiDB does not collect columns of MEDIUMTEXT and LONGTEXT types by default to avoid potential OOM risks.
tidb_enable_historical_stats Modified Changes the default value from ON to OFF, which turns off historical statistics to avoid potential stability issues.
tidb_executor_concurrency Modified Adds support for setting the concurrency of the sort operator.
tidb_sysproc_scan_concurrency Modified Changes the minimum value from 1 to 0. When you set it to 0, TiDB adaptively adjusts the concurrency of scan operations performed when executing internal SQL statements based on the cluster size.
tidb_resource_control_strict_mode Newly added Controls whether privilege control is applied to the SET RESOURCE GROUP statement and the RESOURCE_GROUP() optimizer hint.

Configuration file parameters

Configuration file Configuration parameter Change type Description
TiDB stats-load-concurrency Modified Changes the default value from 5 to 0, and the minimum value from 1 to 0. The value 0 means the automatic mode, which automatically adjusts concurrency based on the configuration of the server.
TiDB token-limit Modified Changes the maximum value from 18446744073709551615 (64-bit platform) and 4294967295 (32-bit platform) to 1048576 to avoid causing TiDB Server OOM when setting it too large. It means that the number of sessions that can execute requests concurrently can be configured to a maximum of 1048576.
TiKV max-apply-unpersisted-log-limit Modified Changes the default value from 0 to 1024 to reduce long-tail latency caused by I/O jitter on the TiKV node. It means that the maximum number of committed but not persisted Raft logs that can be applied is 1024 by default.
TiKV server.grpc-compression-type Modified This configuration item now also controls the compression algorithm of response messages sent from TiKV to TiDB. Enabling compression might consume more CPU resources.
TiFlash security.redact_info_log Modified Introduces a new value option marker. When you set the value to marker, all user data in the log is wrapped in ‹ ›.

System tables

Compiler versions

  • To improve the TiFlash development experience, the minimum version of LLVM required to compile and build TiDB has been upgraded from 13.0 to 17.0. If you are a TiDB developer, you need to upgrade the version of your LLVM compiler to ensure a smooth build. #7193 @Lloyd-Pottiger

Deprecated features

  • The following features are deprecated starting from v8.2.0:

    • Starting from v8.2.0, the enable-replica-selector-v2 configuration item is deprecated. The new version of the Region replica selector is used by default when sending RPC requests to TiKV.
    • Starting from v8.2.0, the BR snapshot restore parameter --concurrency is deprecated. As an alternative, you can configure the maximum number of concurrent tasks per TiKV node during snapshot restore using --tikv-max-restore-concurrency.
    • Starting from v8.2.0, the BR snapshot restore parameter --granularity is deprecated, and the coarse-grained Region scattering algorithm is enabled by default.

  • The following features are planned for deprecation in future versions:

  • The following features are planned for removal in future versions:

    • Starting from v8.0.0, TiDB Lightning deprecates the old version of conflict detection strategy for the physical import mode, and enables you to control the conflict detection strategy for both logical and physical import modes via the conflict.strategy parameter. The duplicate-resolution parameter for the old version of conflict detection will be removed in a future release.

Improvements

  • TiDB

    • Support parallel execution of logical DDL statements (General DDL). Compared with v8.1.0, when you use 10 sessions to submit different DDL statements concurrently, the performance is improved by 3 to 6 times #53246 @D3Hunter
    • Improve the logic of matching multi-column indexes using expressions like ((a = 1 and b = 2 and c > 3) or (a = 4 and b = 5 and c > 6)) and d > 3 to produce a more accurate Range #41598 @ghazalfamilyusa
    • Optimize the performance of obtaining data distribution information when performing simple queries on tables with large data volumes #53850 @you06
    • The aggregated result set can be used as an inner table for IndexJoin, allowing more complex queries to be matched to IndexJoin, thus improving query efficiency through indexing #37068 @elsa0520
    • By batch deleting TiFlash placement rules, improve the processing speed of data GC after performing the TRUNCATE or DROP operation on partitioned tables #54068 @Lloyd-Pottiger
    • Upgrade the version of Azure Identity Libraries and Microsoft Authentication Library to enhance security #53990 @hawkingrei
    • Set the maximum value of token-limit to 1048576 to avoid causing TiDB Server OOM when setting it too large #53312 @djshow832
    • Improve column pruning for MPP execution plans to improve TiFlash MPP execution performance #52133 @yibin87
    • Optimize the performance overhead of the IndexLookUp operator when looking up a table with a large amount of data (>1024 rows) #53871 @crazycs520
    • Remove stores without Regions during MPP load balancing #52313 @xzhangxian1008

  • TiKV

    • Add the Compaction Job Size(files) metric to show the number of SST files involved in a single compaction job #16837 @zhangjinpeng87
    • Enable the early apply feature by default. With this feature enabled, the Raft leader can apply logs after quorum peers have persisted the logs, without waiting for the leader itself to persist the log, reducing the impact of jitter in a few TiKV nodes on write request latency #16717 @glorv
    • Improve the observability of Raft dropped messages to locate the root cause of slow writes #17093 @Connor1996
    • Improve the observability of ingest files latency to troubleshoot cluster latency issues #17078 @LykxSassinator
    • Use a separate thread to clean up Region replicas to ensure stable latency on critical Raft reads and writes #16001 @hbisheng
    • Improve the observability of the number of snapshots being applied #17078 @hbisheng

  • PD

  • TiFlash

    • Reduce lock conflicts under highly concurrent data read operations and optimize short query performance #9125 @JinheLin
    • Eliminate redundant copies of the Join Key in the Join operator #9057 @gengliqi
    • Concurrently perform the process of converting a two-level hash table in the HashAgg operator #8956 @gengliqi
    • Remove redundant aggregation functions for the HashAgg operator to reduce computational overhead #8891 @guo-shaoge

  • Tools

    • Backup & Restore (BR)

      • Optimize the backup feature, improving backup performance and stability during node restarts, cluster scaling-out, and network jitter when backing up large numbers of tables #52534 @3pointer
      • Implement fine-grained checks of TiCDC changefeed during data restore. If the changefeed CheckpointTS is later than the data backup time, the restore operations are not affected, thereby reducing unnecessary wait times and improving user experience #53131 @YuJuncen
      • Add several commonly used parameters to the BACKUP statement and the RESTORE statement, such as CHECKSUM_CONCURRENCY #53040 @RidRisR
      • Except for the br log restore subcommand, all other br log subcommands support skipping the loading of the TiDB domain data structure to reduce memory consumption #52088 @Leavrth
      • Support encryption of temporary files generated during log backup #15083 @YuJuncen
      • Add a tikv_log_backup_pending_initial_scan monitoring metric in the Grafana dashboard #16656 @3pointer
      • Optimize the output format of PITR logs and add a RestoreTS field in the logs #53645 @dveeden

    • TiCDC

      • Support directly outputting raw events when the downstream is a Message Queue (MQ) or cloud storage #11211 @CharlesCheung96

Bug fixes

  • TiDB

    • Fix the issue that when a SQL statement contains an Outer Join and the Join condition includes the false IN (column_name) expression, the query result lacks some data #49476 @ghazalfamilyusa
    • Fix the issue that statistics for columns in system tables are collected when TiDB collects PREDICATE COLUMNS statistics for tables #53403 @hi-rustin
    • Fix the issue that the tidb_enable_column_tracking system variable does not take effect when the tidb_persist_analyze_options system variable is set to OFF #53478 @hi-rustin
    • Fix the issue of potential data races during the execution of (*PointGetPlan).StatsInfo() #49803 #43339 @qw4990
    • Fix the issue that TiDB might return incorrect query results when you query tables with virtual columns in transactions that involve data modification operations #53951 @qw4990
    • Fix the issue that the tidb_enable_async_merge_global_stats and tidb_analyze_partition_concurrency system variables do not take effect during automatic statistics collection #53972 @hi-rustin
    • Fix the issue that TiDB might return the plan not supported error when you query TABLESAMPLE #54015 @tangenta
    • Fix the issue that executing the SELECT DISTINCT CAST(col AS DECIMAL), CAST(col AS SIGNED) FROM ... query might return incorrect results #53726 @hawkingrei
    • Fix the issue that queries cannot be terminated after a data read timeout on the client side #44009 @wshwsh12
    • Fix the overflow issue of the Longlong type in predicates #45783 @hawkingrei
    • Fix the issue that the Window function might panic when there is a related subquery in it #42734 @hi-rustin
    • Fix the issue that the TopN operator might be pushed down incorrectly #37986 @qw4990
    • Fix the issue that SELECT INTO OUTFILE does not work when clustered indexes are used as predicates #42093 @qw4990
    • Fix the issue that the query latency of stale reads increases, caused by information schema cache misses #53428 @crazycs520
    • Fix the issue that comparing a column of YEAR type with an unsigned integer that is out of range causes incorrect results #50235 @qw4990
    • Fix the issue that the histogram and TopN in the primary key column statistics are not loaded after restarting TiDB #37548 @hawkingrei
    • Fix the issue that the final AggMode and the non-final AggMode cannot coexist in Massively Parallel Processing (MPP) #51362 @AilinKid
    • Fix the issue that TiDB panics when executing the SHOW ERRORS statement with a predicate that is always true #46962 @elsa0520
    • Fix the issue that using a view does not work in recursive CTE #49721 @hawkingrei
    • Fix the issue that TiDB might report an error due to GC when loading statistics at startup #53592 @you06
    • Fix the issue that PREPARE/EXECUTE statements with the CONV expression containing a ? argument might result in incorrect query results when executed multiple times #53505 @qw4990
    • Fix the issue that non-BIGINT unsigned integers might produce incorrect results when compared with strings/decimals #41736 @LittleFall
    • Fix the issue that TiDB does not create corresponding statistics metadata (stats_meta) when creating a table with foreign keys #53652 @hawkingrei
    • Fix the issue that certain filter conditions in queries might cause the planner module to report an invalid memory address or nil pointer dereference error #53582 #53580 #53594 #53603 @YangKeao
    • Fix the issue that executing CREATE OR REPLACE VIEW concurrently might result in the table doesn't exist error #53673 @tangenta
    • Fix the issue that the STATE field in the INFORMATION_SCHEMA.TIDB_TRX table is empty due to the size of the STATE field not being defined #53026 @cfzjywxk
    • Fix the issue that the Distinct_count information in GlobalStats might be incorrect when tidb_enable_async_merge_global_stats is disabled #53752 @hawkingrei
    • Fix the issue of incorrect WARNINGS information when using Optimizer Hints #53767 @hawkingrei
    • Fix the issue that negating a time type results in an incorrect value #52262 @solotzg
    • Fix the issue that REGEXP() does not explicitly report an error for empty pattern arguments #53221 @yibin87
    • Fix the issue that converting JSON to datetime might lose precision in some cases #53352 @YangKeao
    • Fix the issue that JSON_QUOTE() returns incorrect results in some cases #37294 @dveeden
    • Fix the issue that executing ALTER TABLE ... REMOVE PARTITIONING might cause data loss #53385 @mjonss
    • Fix the issue that TiDB fails to reject unauthenticated user connections in some cases when using the auth_socket authentication plugin #54031 @lcwangchao
    • Fix the issue that JSON-related functions return errors inconsistent with MySQL in some cases #53799 @dveeden
    • Fix the issue that the INDEX_LENGTH field of partitioned tables in INFORMATION_SCHEMA.PARTITIONS is incorrect #54173 @Defined2014
    • Fix the issue that the TIDB_ROW_ID_SHARDING_INFO field in the INFORMATION_SCHEMA.TABLES table is incorrect #52330 @tangenta
    • Fix the issue that a generated column returns illegal timestamps #52509 @lcwangchao
    • Fix the issue that setting max-index-length causes TiDB to panic when adding indexes using the Distributed eXecution Framework (DXF) #53281 @zimulala
    • Fix the issue that the illegal column type DECIMAL(0,0) can be created in some cases #53779 @tangenta
    • Fix the issue that using CURRENT_DATE() as the default value for a column results in incorrect query results #53746 @tangenta
    • Fix the issue that the ALTER DATABASE ... SET TIFLASH REPLICA statement incorrectly adds TiFlash replicas to the SEQUENCE table #51990 @jiyfhust
    • Fix the issue that the REFERENCED_TABLE_SCHEMA field in the INFORMATION_SCHEMA.KEY_COLUMN_USAGE table is incorrect #52350 @wd0517
    • Fix the issue that inserting multiple rows in a single statement causes the AUTO_INCREMENT column to be discontinuous when AUTO_ID_CACHE=1 #52465 @tiancaiamao
    • Fix the format of deprecation warnings #52515 @dveeden
    • Fix the issue that the TRACE command is missing in copr.buildCopTasks #53085 @time-and-fate
    • Fix the issue that the memory_quota hint might not work in subqueries #53834 @qw4990
    • Fix the issue that improper use of metadata locks might lead to writing anomalous data when using the plan cache under certain circumstances #53634 @zimulala
    • Fix the issue that after a statement within a transaction is killed by OOM, if TiDB continues to execute the next statement within the same transaction, you might get an error Trying to start aggressive locking while it's already started and a panic occurs #53540 @MyonKeminta

  • TiKV

    • Fix the issue that pushing down the JSON_ARRAY_APPEND() function to TiKV causes TiKV to panic #16930 @dbsid
    • Fix the issue that the leader does not clean up failed snapshot files in time #16976 @hbisheng
    • Fix the issue that highly concurrent Coprocessor requests might cause TiKV OOM #16653 @overvenus
    • Fix the issue that changing the raftstore.periodic-full-compact-start-times configuration item online might cause TiKV to panic #17066 @SpadeA-Tang
    • Fix the failure of make docker and make docker_test #17075 @shunki-fujita
    • Fix the issue that the gRPC request sources duration metric is displayed incorrectly in the monitoring dashboard #17133 @King-Dylan
    • Fix the issue that setting the gRPC message compression method via grpc-compression-type does not take effect on messages sent from TiKV to TiDB #17176 @ekexium
    • Fix the issue that the output of the raft region command in tikv-ctl does not include the Region status information #17037 @glorv
    • Fix the issue that CDC and log-backup do not limit the timeout of check_leader using the advance-ts-interval configuration, causing the resolved_ts lag to be too large when TiKV restarts normally in some cases #17107 @MyonKeminta

  • PD

    • Fix the issue that ALTER PLACEMENT POLICY cannot modify the placement policy #52257 #51712 @jiyfhust
    • Fix the issue that the scheduling of write hotspots might break placement policy constraints #7848 @lhy1024
    • Fix the issue that down peers might not recover when using Placement Rules #7808 @rleungx
    • Fix the issue that a large number of retries occur when canceling resource groups queries #8217 @nolouch
    • Fix the issue that manually transferring the PD leader might fail #8225 @HuSharp

  • TiFlash

    • Fix the issue of query timeout when executing queries on partitioned tables that contain empty partitions #9024 @JinheLin
    • Fix the issue that in the disaggregated storage and compute architecture, null values might be incorrectly returned in queries after adding non-null columns in DDL operations #9084 @Lloyd-Pottiger
    • Fix the issue that the SUBSTRING_INDEX() function might cause TiFlash to crash in some corner cases #9116 @wshwsh12
    • Fix the issue that a large number of duplicate rows might be read in FastScan mode after importing data via BR or TiDB Lightning #9118 @JinheLin

  • Tools

    • Backup & Restore (BR)

      • Fix the issue that BR fails to restore a transactional KV cluster due to an empty EndKey #52574 @3pointer
      • Fix the issue that a PD connection failure could cause the TiDB instance where the log backup advancer owner is located to panic #52597 @YuJuncen
      • Fix the issue that log backup might be paused after the advancer owner migration #53561 @RidRisR
      • Fix the issue that BR fails to correctly identify errors due to multiple nested retries during the restore process #54053 @RidRisR
      • Fix the issue that the connection used to fetch TiKV configurations might not be closed #52595 @RidRisR
      • Fix the issue that the TestStoreRemoved test case is unstable #52791 @YuJuncen
      • Fix the issue that TiFlash crashes during point-in-time recovery (PITR) #52628 @RidRisR
      • Fix the inefficiency issue in scanning DDL jobs during incremental backups #54139 @3pointer
      • Fix the issue that the backup performance during checkpoint backups is affected due to interruptions in seeking Region leaders #17168 @Leavrth

    • TiCDC

      • Fix inaccurate display of the Kafka Outgoing Bytes panel in Grafana #10777 @asddongmen
      • Fix the issue that data inconsistency might occur when restarting Changefeed repeatedly when performing a large number of UPDATE operations in a multi-node environment #11219 @lidezhu

    • TiDB Data Migration (DM)

      • Fix the connection blocking issue by upgrading go-mysql #11041 @D3Hunter
      • Fix the issue that SET statements cause DM to panic during the migration of MariaDB data #10206 @dveeden

    • TiDB Lightning

      • Fix the issue that TiDB Lightning might report an error when importing zstd compressed files #53587 @lance6716

    • Dumpling

      • Fix the issue that Dumpling reports an error when exporting tables and views at the same time #53682 @tangenta

    • TiDB Binlog

      • Fix the issue that deleting rows during the execution of ADD COLUMN might report an error data and columnID count not match when TiDB Binlog is enabled #53133 @tangenta

Contributors

We would like to thank the following contributors from the TiDB community: