Important

You are viewing the archived documentation of TiDB, which no longer receives updates. It is recommended that you use the latest LTS version of the TiDB database.

TiDB Backend Task Distributed Execution Framework

Warning

This feature is an experimental feature. It is not recommended to use it in production environments.

TiDB adopts a computing-storage separation architecture with excellent scalability and elasticity. Starting from v7.1.0, TiDB introduces a backend task distributed execution framework to further leverage the resource advantages of the distributed architecture. The goal of this framework is to implement unified scheduling and distributed execution of all backend tasks, and to provide unified resource management capabilities for both overall and individual backend tasks, which better meets users' expectations for resource usage.

This document describes the use cases, limitations, usage, and implementation principles of the TiDB backend task distributed execution framework.

Note

This framework does not support the distributed execution of SQL queries.

Use cases and limitations

In a database management system, in addition to the core transactional processing (TP) and analytical processing (AP) workloads, there are other important tasks, such as DDL operations, IMPORT INTO, TTL, Analyze, and Backup/Restore, which are called backend tasks. These backend tasks need to process a large amount of data in database objects (tables), so they typically have the following characteristics:

Need to process all data in a schema or a database object (table).
Might need to be executed periodically, but at a low frequency.
If the resources are not properly controlled, they are prone to affect TP and AP tasks, lowering the database service quality.

Enabling the TiDB backend task distributed execution framework can solve the above problems and has the following three advantages:

The framework provides unified capabilities for high scalability, high availability, and high performance.
The framework supports distributed execution of backend tasks, which can flexibly schedule the available computing resources of the entire TiDB cluster, thereby better utilizing the computing resources in a TiDB cluster.
The framework provides unified resource usage and management capabilities for both overall and individual backend tasks.

Currently, for TiDB Self-Hosted, the TiDB backend task distributed execution framework supports the distributed execution of the ADD INDEX and IMPORT INTO statements. For TiDB Cloud, the IMPORT INTO statement is not applicable.

ADD INDEX is a DDL statement used to create indexes. For example:

ALTER TABLE t1 ADD INDEX idx1(c1);
CREATE INDEX idx1 ON table t1(c1);

IMPORT INTO is used to import data in formats such as CSV, SQL, and PARQUET into an empty table. For more information, see IMPORT INTO.

Limitation

The DXF can only schedule the distributed execution of one ADD INDEX task at a time. If a new ADD INDEX task is submitted before the current ADD INDEX distributed task has finished, the new task is executed through a transaction.
Adding indexes on columns with the TIMESTAMP data type through the DXF is not supported, because it might lead to inconsistency between the index and the data.

Prerequisites

Before using the distributed framework, you need to enable the Fast Online DDL mode.

Adjust the following system variables related to Fast Online DDL:
- tidb_ddl_enable_fast_reorg: used to enable Fast Online DDL mode. It is enabled by default starting from TiDB v6.5.0.
- tidb_ddl_disk_quota: used to control the maximum quota of local disks that can be used in Fast Online DDL mode.
Adjust the following configuration item related to Fast Online DDL:
- temp-dir: specifies the local disk path that can be used in Fast Online DDL mode.

Note

Before you upgrade TiDB to v6.5.0 or later, it is recommended that you check whether the temp-dir path of TiDB is correctly mounted to an SSD disk. Make sure that the operating system user that runs TiDB has the read and write permissions for this directory. Otherwise, The DDL operations might experience unpredictable issues. This path is a TiDB configuration item, which takes effect after TiDB is restarted. Therefore, setting this configuration item in advance before upgrading can avoid another restart.

Usage

To enable the distributed framework, set the value of tidb_enable_dist_task to ON:
```
SET GLOBAL tidb_enable_dist_task = ON;
```
When background tasks are running, the statements supported by the framework (such as ADD INDEX and IMPORT INTO) are executed in a distributed manner. All TiDB nodes run background tasks by default.
Adjust the following system variables that might affect the distributed execution of DDL tasks according to your needs:
- tidb_ddl_reorg_worker_cnt: use the default value 4. The recommended maximum value is 16.
- tidb_ddl_reorg_priority
- tidb_ddl_error_count_limit
- tidb_ddl_reorg_batch_size: use the default value. The recommended maximum value is 1024.
Starting from v7.4.0, you can adjust the number of TiDB nodes that perform background tasks according to actual needs. After deploying a TiDB cluster, you can set the instance-level system variable tidb_service_scope for each TiDB node in the cluster. When tidb_service_scope of a TiDB node is set to background, the TiDB node can execute background tasks. When tidb_service_scope of a TiDB node is set to the default value "", the TiDB node cannot execute background tasks. If tidb_service_scope is not set for any TiDB node in a cluster, the TiDB distributed execution framework schedules all TiDB nodes to execute background tasks by default.
Warning
tidb_service_scope is an experimental feature. It is not recommended to use it in production environments. This feature might be changed or removed without prior notice. If you find a bug, you can report an issue on GitHub.

Tip

For distributed execution of ADD INDEX statements, you only need to set tidb_ddl_reorg_worker_cnt.

Implementation principles

The architecture of the TiDB backend task distributed execution framework is as follows:

Architecture of the TiDB backend task distributed execution framework

As shown in the preceding diagram, the execution of backend tasks in the distributed framework is mainly handled by the following modules:

Dispatcher: generates the distributed execution plan for each task, manages the execution process, converts the task status, and collects and feeds back the runtime task information.
Scheduler: replicates the execution of distributed tasks among TiDB nodes to improve the efficiency of backend task execution.
Subtask Executor: the actual executor of distributed subtasks. In addition, the Subtask Executor returns the execution status of subtasks to the Scheduler, and the Scheduler updates the execution status of subtasks in a unified manner.
Resource pool: provides the basis for quantifying resource usage and management by pooling computing resources of the above modules.