TiDB 8.5.6 Release Notes

May 21, 2026 ยท View on GitHub

Release date: April 14, 2026

TiDB version: 8.5.6

Quick access: Quick start | Production deployment

Features

Performance

  • Foreign key checks now support shared locks #66154 @you06

    In pessimistic transactions, when you run INSERT or UPDATE on a child table with foreign key constraints, foreign key checks lock the corresponding parent table rows with exclusive locks by default. In high-concurrency write scenarios on the child table, if many transactions access the same parent table rows, severe lock contention can occur.

    Starting from v8.5.6, you can set the tidb_foreign_key_check_in_shared_lock system variable to ON to let foreign key checks use shared locks on the parent table, thereby reducing lock contention and improving concurrent write performance on the child table.

    For more information, see documentation.

Stability

  • The feature of setting the maximum limit on resource usage for background tasks of resource control becomes generally available (GA) #56019 @glorv

    TiDB resource control can identify and lower the priority of background tasks. In certain scenarios, you might want to limit the resource consumption of background tasks, even when resources are available. Starting from v8.4.0, you can use the UTILIZATION_LIMIT parameter to set the maximum percentage of resources that background tasks can consume. Each node will keep the resource usage of all background tasks below this percentage. This feature enables precise control over resource consumption for background tasks, further enhancing cluster stability.

    In v8.5.6, this feature is generally available (GA).

    For more information, see documentation.

Observability

  • Support defining multi-dimensional, fine-grained trigger rules for slow query logs #62959, #64010 @zimulala

    Before v8.5.6, the main way to identify slow queries in TiDB is to set the tidb_slow_log_threshold system variable. This mechanism provides only coarse-grained control over slow query log triggering because it applies globally at the instance level and does not support fine-grained control at the session or SQL level. In addition, it supports only one trigger condition, execution time (Query_time), which cannot meet the need to capture slow query logs more precisely in complex scenarios.

    Starting from v8.5.6, TiDB enhances slow query log control. You can use the tidb_slow_log_rules system variable to define multi-dimensional slow query log output rules at the instance, session, and SQL levels, based on conditions such as Query_time, Digest, Mem_max, and KV_total. You can use tidb_slow_log_max_per_sec to limit the number of log entries written per second, and use the WRITE_SLOW_LOG hint to force slow query logging for specific SQL statements. This enables more flexible and fine-grained control over slow query logs.

    For more information, see documentation.

SQL

  • Support column-level privilege management #61706 @CbcWestwolf @fzzf678

    Before v8.5.6, TiDB privilege control covers the database and table levels and does not support granting or revoking privileges on specific columns, unlike MySQL. As a result, you cannot restrict users to access only a subset of sensitive columns in a table.

    Starting from v8.5.6, TiDB supports column-level privilege management. You can use the GRANT and REVOKE statements to manage privileges on specific columns. TiDB performs privilege checks based on column-level privileges during query processing and execution plan construction, enabling finer-grained access control and better support for sensitive data isolation and the principle of least privilege.

    For more information, see documentation.

  • Support using table aliases in the FOR UPDATE OF clause #63035 @cryo-zd

    Before v8.5.6, when a SELECT ... FOR UPDATE OF <table> statement references a table alias in the locking clause, TiDB might fail to resolve the alias correctly and return the table not exists error even if the alias is valid.

    Starting from v8.5.6, TiDB supports using table aliases in the FOR UPDATE OF clause. TiDB can now correctly resolve locking targets from the FROM clause, including aliased tables, ensuring that row locks take effect as expected. This improves MySQL compatibility and makes SELECT ... FOR UPDATE OF statements more stable and reliable in queries that use table aliases.

    For more information, see documentation.

DB operations

  • Support specifying the maximum number of nodes for Distributed eXecution Framework (DXF) tasks #58944 @tangenta @D3Hunter

    Before v8.5.6, TiDB does not provide a way to limit the number of nodes used by a distributed execution task. When you want to control resource usage for DXF, TiDB does not provide a dedicated option to constrain the maximum node count.

    Starting from v8.5.6, TiDB introduces the tidb_max_dist_task_nodes system variable to specify the maximum number of TiDB nodes used by a DXF task, enabling better resource control and workload-based tuning.

    For more information, see documentation.

Data Migration

  • Migrate sync-diff-inspector from pingcap/tidb-tools to pingcap/tiflow repository #11672 @joechenrh

Compatibility changes

For TiDB clusters newly deployed in v8.5.5 (that is, not upgraded from versions earlier than v8.5.4), you can smoothly upgrade to v8.5.6. Most changes in v8.5.6 are safe for routine upgrades, but this release also includes several MySQL compatibility changes, system variable updates, configuration parameter updates, and deprecated features. Before upgrading, make sure to carefully review this section.

MySQL compatibility

  • Starting from v8.5.6, TiDB supports a MySQL-compatible column-level privilege management mechanism. You can grant or revoke SELECT, INSERT, UPDATE, and REFERENCES privileges for specific columns at the table level. For more information, see Column-Level Privilege Management.
  • Starting from v8.5.6, TiDB supports using table aliases in the FOR UPDATE OF clause. To maintain backward compatibility, you can still reference the base table name when an alias is defined, but this triggers a warning recommending the use of an explicit alias. For more information, see SELECT.
  • Starting from v8.5.6, Dumpling supports exporting data from MySQL 8.4 by adopting the updated MySQL binary log terminology. #53082 @dveeden
  • Starting from v8.5.6, TiDB Data Migration (DM) supports MySQL 8.4 as an upstream data source by adapting to the new terminology and version detection logic introduced in this version. #11020 @dveeden

System variables

Variable nameChange typeDescription
tidb_analyze_versionModifiedStarting from v8.5.6, statistics Version 1 (tidb_analyze_version = 1) is deprecated and will be removed in a future release. It is recommended to use statistics Version 2 (tidb_analyze_version = 2).
tidb_ignore_inlist_plan_digestModifiedChanges the default value from OFF to ON. The default value ON means that TiDB ignores the element differences (including the difference in the number of elements) in the IN list and uses ... to replace elements in the IN list when generating Plan Digests.
tidb_service_scopeModifiedStarting from v8.5.6, the value of this variable is case-insensitive. TiDB converts the input value to lowercase for storage and comparison.
InPacketBytesNewly addedThis variable is used only for internal statistics and is not visible to users.
OutPacketBytesNewly addedThis variable is used only for internal statistics and is not visible to users.
tidb_foreign_key_check_in_shared_lockNewly addedControls whether foreign key checks in pessimistic transactions use shared locks instead of exclusive locks on rows in the parent table. The default value is OFF, which means TiDB uses exclusive locks by default.
tidb_max_dist_task_nodesNewly addedDefines the maximum number of TiDB nodes that the Distributed eXecution Framework (DXF) tasks can use. The default value is -1, which indicates that automatic mode is enabled. In automatic mode, TiDB dynamically calculates the value as min(3, tikv_nodes / 3), where tikv_nodes represents the number of TiKV nodes in the cluster.
tidb_opt_join_reorder_through_selNewly addedImproves join reorder optimization for certain multi-table join queries. If you set it to ON and safety conditions are met, the optimizer evaluates Selection conditions between consecutive join operators together with join order candidates. During join tree reconstruction, the optimizer pushes these conditions down to more appropriate positions whenever possible, allowing more tables to participate in join order optimization.
tidb_opt_partial_ordered_index_for_topnNewly addedControls whether the optimizer can leverage the partial ordering of an index to optimize TopN computation when a query contains ORDER BY ... LIMIT. The default value is DISABLE, which means the optimization is disabled.
tidb_slow_log_max_per_secNewly addedControls the maximum number of slow query log entries that can be written per TiDB node per second.
  • A value of 0 (the default) means there is no limit on the number of slow query log entries written per second.
  • A value greater than 0 means TiDB writes at most the specified number of slow query log entries per second. Any excess log entries are discarded and not written to the slow query log file.
tidb_slow_log_rulesNewly addedDefines the triggering rules for slow query logs. It supports combining multi-dimensional metrics to provide more flexible and fine-grained logging.

Configuration parameters

Configuration file or componentConfiguration parameterChange typeDescription
TiKVgc.auto-compaction.mvcc-read-aware-enabledNewly addedControls whether to enable MVCC-read-aware compaction. The default value is false.
TiKVgc.auto-compaction.mvcc-read-weightNewly addedThe weight multiplier applied to MVCC read activity when calculating the compaction priority score for a Region. The default value is 3.0.
TiKVgc.auto-compaction.mvcc-scan-thresholdNewly addedThe minimum number of MVCC versions scanned per read request to mark a Region as a compaction candidate. The default value is 1000.
TiCDCsink.csv.output-field-headerNewly addedControls whether a header row is output in CSV files. The default value is false. This parameter applies only to the TiCDC new architecture.

System table changes

System tableChange typeDescription
mysql.tidbModifiedAdds the cluster_id field, which represents the unique identifier of a TiDB cluster. Note that cluster_id is read-only and cannot be modified.

Deprecated features

  • Starting from v8.5.6, statistics Version 1 (tidb_analyze_version = 1) is deprecated and will be removed in a future release. It is recommended that you use statistics Version 2 (tidb_analyze_version = 2) and migrate existing objects that use statistics Version 1 to Version 2 for more accurate statistics.
  • Starting from v8.5.6, the TiDB Lightning Web Interface is deprecated and will be removed in v8.5.7. The web UI build has been broken since v8.4.0. Use the CLI or the IMPORT INTO statement instead. If this affects your workflow, comment on #67697.

Improvements

  • TiDB

    • Improve plan selection for queries with IN predicates on index prefix columns. TiDB can now use merge sort to preserve order for ORDER BY ... LIMIT queries, reducing unnecessary scans and improving performance. #63449 #34882 @time-and-fate
    • Improve slow query log readability by outputting non-printable prepared statement arguments as hexadecimal values #65383 @dveeden
    • Add cluster_id to mysql.tidb, enabling external tools to determine whether two TiDB instances belong to the same cluster #59476 @YangKeao

  • TiKV

    • Introduce a load-based compaction mechanism, which detects MVCC read overhead and prioritizes compaction for Regions with higher read cost to improve query performance #19133 @mittalrishabh
    • Optimize the stale range cleanup logic during cluster scale-out and scale-in operations by deleting stale keys directly instead of cleaning them up through SST file ingestion, thereby reducing the impact on online request latency #18042 @LykxSassinator

  • PD

  • Tools

    • TiDB Data Migration (DM)

      • Add foreign key causal dependency support in DM syncer to ensure parent-to-child execution order of row changes in multi-worker scenarios #12350 @OliverS929

Bug fixes

  • TiDB

    • Fix the issue that upgrading from release-8.5-20250606-v8.5.2 to the upstream release-8.5 might skip the PITR metadata upgrade and cause PITR operations to fail #66994 @fzzf678
    • Fix the issue that after executing EXCHANGE PARTITION, non-unique global indexes or nullable unique global indexes on non-clustered partitioned tables might become inconsistent and return incomplete results #65289 @mjonss
    • Fix the issue that KILL QUERY might incorrectly terminate idle connections #65447 @gengliqi
    • Fix the issue that the value of tidb_service_scope is not converted to lowercase when set #66749 @D3Hunter
    • Fix the issue that affinity tables are not displayed after TiDB restarts #66284 @lcwangchao
    • Fix the issue that the Stats Healthy metric displays inaccurately because system tables are not excluded from the stats cache #64080 @0xPoe
    • Fix the issue that statistics might not be updated due to abnormal updates of modify_count #65426 @0xPoe
    • Fix the issue that a pessimistic transaction might be unexpectedly rolled back due to a keep-alive mechanism failure when its first statement acquires locks in fair locking mode #66571 @MyonKeminta

  • TiKV

    • Fix the memory leak issue in crossbeam skiplist #19285 @ekexium
    • Fix the issue that global indexes on non-unique columns of partitioned tables might become inconsistent and return incorrect results in some cases #19262 @mjonss
    • Fix the issue that stalled coprocessor snapshot retrieval might occupy unified read pool workers until request deadlines expire, delaying other read requests #18491 @AndreMouche
    • Fix the issue that follower reads might remain blocked on disk-full TiKV nodes by rejecting read-index requests on disk-full followers #19201 @glorv
    • Fix the issue that resolved-ts task backlogs might cause OOM when the resolved-ts worker is busy #18359 @overvenus
    • Fix the issue that long-tail follower read latency might occur during leader transfer by retrying read-index requests earlier and adding a dedicated retry interval setting #18417 @gengliqi
    • Fix the rare data inconsistency issue when retrying a prewrite request in pessimistic transactions #11187 @wk989898

  • PD

    • Fix a panic issue that might occur when executing DISTRIBUTE TABLE in scenarios with a large number of Merge Region operators #10293 @bufferflies
    • Fix the issue that configuring Store Limit might not take effect immediately #10108 @okJiang

  • TiFlash

    • Fix a potential data inconsistency issue between TiFlash and TiKV after executing a DDL statement to remove the NOT NULL constraint of a column #10680 @JaySon-Huang
    • Fix the issue that the Raft throughput metric in the Grafana dashboard might display an abnormally large value #10701 @CalvinNeo
    • Fix the issue that the Join result might be incorrect when the Runtime Filter is enabled and the Join Keys have inconsistent data types #10699 @ChangRui-Ryan

  • Tools

    • Backup & Restore (BR)

      • Fix the issue that flush_ts might be 0 in log backup #19406 @YuJuncen
      • Fix the issue that BR might fail during multipart uploads when accessing Google Cloud Storage via Amazon S3-compatible APIs with S3-style credentials, due to a missing Content-Length header #19352 @Leavrth
      • Fix the issue that the BR restore point might remain stuck in the waiting for schema info finishes reloading state for a long time and then fail due to timeout after 15 minutes #66110 @kennytm
      • Fix the issue that BR cannot correctly pre-split Regions when restoring tables with the SHARD_ROW_ID_BITS, PRE_SPLIT_REGIONS, and merge_option attributes #65060 @JoyC-dev

    • TiCDC

      • Fix the issue that changefeeds might repeatedly create invalid dispatchers after the server restarts #4452 @wlwilliamx
      • Fix the issue that TiCDC fails to replicate table renaming operations correctly when the upstream TiDB version is v8.1.x or earlier #4392 @lidezhu
      • Fix the issue that TiKV might crash during data scanning when TiCDC is enabled #19404 @wk989898
      • Support Azure Managed Identity authentication for Azure Blob Storage and fix the issue that uploads to cloud storage might get stuck #3093 @wlwilliamx

    • TiDB Data Migration (DM)

      • Fix the issue that DM does not advance the global checkpoint position after an upstream binlog file rotation #12339 @OliverS929
      • Fix the issue that, when processing updates on tables with foreign key constraints in safe mode, DM might still incorrectly trigger foreign key cascades and cause unintended data deletion even if the primary key or unique key is not modified #12350 @OliverS929
      • Fix the issue that DM validator incorrectly returns validation errors when processing UNSIGNED columns #12178 @OliverS929