TiDB Data Migration Table Routing

When you migrate data using TiDB Data Migration (DM), you can configure the table routing to migrate a certain table of the upstream MySQL or MariaDB instance to the specified table in the downstream.

Table Routing - 图1

Note

  • Configuring multiple different routing rules for a single table is not supported.
  • The match rule of schema needs to be configured separately, which is used to migrate CREATE/DROP SCHEMA xx, as shown in rule-2 of the Configure table routing section.

Configure table routing

  1. routes:
  2.   rule-1:
  3.     schema-pattern: "test_*"
  4.     table-pattern: "t_*"
  5.     target-schema: "test"
  6.     target-table: "t"
  7.     # extract-table, extract-schema, and extract-source are optional and
  8.     # are required only when you need to extract information about sharded
  9.     # tables, sharded schemas, and source datatabase information.
  10.     extract-table:
  11.       table-regexp: "t_(.*)"
  12.       target-column: "c_table"
  13.     extract-schema:
  14.       schema-regexp: "test_(.*)"
  15.       target-column: "c_schema"
  16.     extract-source:
  17.       source-regexp: "(.*)"
  18.       target-column: "c_source"
  19.   rule-2:
  20.     schema-pattern: "test_*"
  21.     target-schema: "test"

Regular expressions and wildcards are supported to match database and table names. In simple scenarios, it is recommended that you use the wildcard for matching schemas and tables. However, note the following:

  • Wildcards including *, ?, and [] are supported. There can only be one * symbol in a wildcard match, and it must be at the end. For example, in table-pattern: "t_*", "t_*" indicates all tables starting with t_. See wildcard matching#Syntax) for details.

  • table-regexp, schema-regexp, and source-regexp only support regular expressions and cannot start with the ~ symbol.

  • schema-pattern and table-pattern support both wildcards and regular expressions. Regular expressions must begin with the ~ symbol.

Parameter descriptions

  • DM migrates the upstream MySQL or MariaDB instance tables that match the schema-pattern/table-pattern rule provided by Table selector to the downstream target-schema/target-table.
  • For sharded tables that match the schema-pattern/table-pattern rules, DM extracts the table name by using the extract-table.table-regexp regular expression, the schema name by using the extract-schema.schema-regexp regular expression, and source information by using the extract-source.source-regexp regular expression. Then DM writes the extracted information to the corresponding target-column in the merged table in the downstream.

Usage examples

This section shows the usage examples in four different scenarios.

If you need to migrate and merge MySQL shards of small datasets to TiDB, refer to this tutorial.

Merge sharded schemas and tables

Assuming in the scenario of sharded schemas and tables, you want to migrate the test_{1,2,3...}.t_{1,2,3...} tables in two upstream MySQL instances to the test.t table in the downstream TiDB instance.

To migrate the upstream instances to the downstream test.t, you must create the following routing rules:

  • rule-1 is used to migrate DML or DDL statements of the table that matches schema-pattern: "test_*" and table-pattern: "t_*" to the downstream test.t.
  • rule-2 is used to migrate DDL statements of the schema that matches schema-pattern: "test_*", such as CREATE/DROP SCHEMA xx.

Table Routing - 图2

Note

  • If the downstream schema: test already exists and is not to be deleted, you can omit rule-2.
  • If the downstream schema: test does not exist and only rule-1 is configured, then it reports the schema test doesn't exist error during migration.
  1.   rule-1:
  2.     schema-pattern: "test_*"
  3.     table-pattern: "t_*"
  4.     target-schema: "test"
  5.     target-table: "t"
  6.   rule-2:
  7.     schema-pattern: "test_*"
  8.     target-schema: "test"

Extract table, schema, and source information and write into the merged table

Assuming in the scenario of sharded schemas and tables, you want to migrate the test_{1,2,3...}.t_{1,2,3...} tables in two upstream MySQL instances to the test.t table in the downstream TiDB instance. At the same time, you want to extract the source information of the sharded tables and write it to the downstream merged table.

To migrate the upstream instances to the downstream test.t, you must create routing rules similar to the previous section Merge sharded schemas and tables. In addtion, you need to add the extract-table, extract-schema, and extract-source configurations:

  • extract-table: For a sharded table matching schema-pattern and table-pattern, DM extracts the sharded table name by using table-regexp and writes the name suffix without the t_ part to target-column of the merged table, that is, the c_table column.
  • extract-schema: For a sharded schema matching schema-pattern and table-pattern, DM extracts the sharded schema name by using schema-regexp and writes the name suffix without the test_ part to target-column of the merged table, that is, the c_schema column.
  • extract-source: For a sharded table matching schema-pattern and table-pattern, DM writes the source instance information to the target-column of the merged table, that is, the c_source column.
  1.   rule-1:
  2.     schema-pattern: "test_*"
  3.     table-pattern: "t_*"
  4.     target-schema: "test"
  5.     target-table: "t"
  6.     extract-table:
  7.       table-regexp: "t_(.*)"
  8.       target-column: "c_table"
  9.     extract-schema:
  10.       schema-regexp: "test_(.*)"
  11.       target-column: "c_schema"
  12.     extract-source:
  13.       source-regexp: "(.*)"
  14.       target-column: "c_source"
  15.   rule-2:
  16.     schema-pattern: "test_*"
  17.     target-schema: "test"

To extract the source information of upstream sharded tables to the merged table in the downstream, you must manually create a merged table in the downstream before starting the migration. The merged table must contain the three target-columns (c_table, c_schema, and c_source) used for specifying the source information. In addition, these columns must be the last columns and be string types.

  1. CREATE TABLE `test`.`t` (
  2.     a int(11) PRIMARY KEY,
  3.     c_table varchar(10) DEFAULT NULL,
  4.     c_schema varchar(10) DEFAULT NULL,
  5.     c_source varchar(10) DEFAULT NULL
  6. );

Assume that the upstream has the following two data sources:

Data source mysql-01:

  1. mysql> select * from test_11.t_1;
  2. +---+
  3. | a |
  4. +---+
  5. | 1 |
  6. +---+
  7. mysql> select * from test_11.t_2;
  8. +---+
  9. | a |
  10. +---+
  11. | 2 |
  12. +---+
  13. mysql> select * from test_12.t_1;
  14. +---+
  15. | a |
  16. +---+
  17. | 3 |
  18. +---+

Data source mysql-02:

  1. mysql> select * from test_13.t_3;
  2. +---+
  3. | a |
  4. +---+
  5. | 4 |
  6. +---+

After migration using DM, data in the merged table will be as follows:

  1. mysql> select * from test.t;
  2. +---+---------+----------+----------+
  3. | a | c_table | c_schema | c_source |
  4. +---+---------+----------+----------+
  5. | 1 | 1       | 11       | mysql-01 |
  6. | 2 | 2       | 11       | mysql-01 |
  7. | 3 | 1       | 12       | mysql-01 |
  8. | 4 | 3       | 13       | mysql-02 |
  9. +---+---------+----------+----------+

Incorrect examples of creating merged tables

Table Routing - 图3

Note

If any of the following errors occur, source information of sharded tables and schemas might fail to be written to the merged table.

  • c-table is not in the last three columns:
  1. CREATE TABLE `test`.`t` (
  2.     c_table varchar(10) DEFAULT NULL,
  3.     a int(11) PRIMARY KEY,
  4.     c_schema varchar(10) DEFAULT NULL,
  5.     c_source varchar(10) DEFAULT NULL
  6. );
  • c-source is absent:
  1. CREATE TABLE `test`.`t` (
  2.     a int(11) PRIMARY KEY,
  3.     c_table varchar(10) DEFAULT NULL,
  4.     c_schema varchar(10) DEFAULT NULL,
  5. );
  • c_schema is not a string type:
  1. CREATE TABLE `test`.`t` (
  2.     a int(11) PRIMARY KEY,
  3.     c_table varchar(10) DEFAULT NULL,
  4.     c_schema int(11) DEFAULT NULL,
  5.     c_source varchar(10) DEFAULT NULL,
  6. );

Merge sharded schemas

Assuming in the scenario of sharded schemas, you want to migrate the test_{1,2,3...}.t_{1,2,3...} tables in the two upstream MySQL instances to the test.t_{1,2,3...} tables in the downstream TiDB instance.

To migrate the upstream schemas to the downstream test.t_[1,2,3], you only need to create one routing rule.

  1.   rule-1:
  2.     schema-pattern: "test_*"
  3.     target-schema: "test"

Incorrect table routing

Assuming that the following two routing rules are configured and test_1_bak.t_1_bak matches both rule-1 and rule-2, an error is reported because the table routing configuration violates the number limitation.

  1.   rule-1:
  2.     schema-pattern: "test_*"
  3.     table-pattern: "t_*"
  4.     target-schema: "test"
  5.     target-table: "t"
  6.   rule-2:
  7.     schema-pattern: "test_1_bak"
  8.     table-pattern: "t_1_bak"
  9.     target-schema: "test"
  10.     target-table: "t_bak"