ddl, meta: support altering auto_increment ID to a smaller value

[tangenta]

What problem does this PR solve?

Issue Number: close #25400

Problem Summary:

For now, the auto ID in TiDB grows only and can never bet set 'back'.

If the auto ID is rebased to a very large value because of either mistake or bug, it is almost impossible to set it back, this can be a headache for the users.

For details, see #25400.

What is changed and how it works?

What's Changed:

This PR introduces a new syntax to force changing the next global auto ID:

ALTER TABLE `t` FORCE AUTO_INCREMENT = 3;
ALTER TABLE `t` /*T![force_auto_inc] FORCE */ AUTO_INCREMENT = 3;

drop table if exists t;
create table t (a bigint primary key auto_increment);
alter table t FORCE auto_increment = 18446744073709551615;
show table t next_row_id;
insert into t values ();
alter table t FORCE auto_increment = 10086;
show table t next_row_id;
insert into t values();
select a from t;

mysql> alter table t FORCE auto_increment = 18446744073709551615;
Query OK, 0 rows affected (0.01 sec)

mysql> show table t next_row_id;
+---------+------------+-------------+----------------------+----------------+
| DB_NAME | TABLE_NAME | COLUMN_NAME | NEXT_GLOBAL_ROW_ID   | ID_TYPE        |
+---------+------------+-------------+----------------------+----------------+
| test    | t          | a           | 18446744073709551615 | AUTO_INCREMENT |
+---------+------------+-------------+----------------------+----------------+
1 row in set (0.00 sec)

mysql> insert into t values ();
ERROR 1467 (HY000): Failed to read auto-increment value from storage engine

mysql> alter table t FORCE auto_increment = 10086;
Query OK, 0 rows affected (0.01 sec)

mysql> show table t next_row_id;
+---------+------------+-------------+--------------------+----------------+
| DB_NAME | TABLE_NAME | COLUMN_NAME | NEXT_GLOBAL_ROW_ID | ID_TYPE        |
+---------+------------+-------------+--------------------+----------------+
| test    | t          | a           |              10086 | AUTO_INCREMENT |
+---------+------------+-------------+--------------------+----------------+
1 row in set (0.00 sec)

mysql> insert into t values();
Query OK, 1 row affected (0.00 sec)

mysql> select a from t;
+-------+
| a     |
+-------+
| 10086 |
+-------+
1 row in set (0.00 sec)

Please attention: the FORCE keyword should be only used to recover metadata. All the write workload on the target table should be stopped and the ID should be the largest existing ID plus one(select max(id)+1 from t). Otherwise, it may lose row data.

Related changes

• PR to update pingcap/docs/pingcap/docs-cn:
• Need to cherry-pick to the release branch?

Check List

Tests

• Unit test
• Integration test

Side effects

NA

Release note

• Supports forcing the next auto-increment ID to be altered.

[AilinKid]

LGTM

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• AilinKid
• bb7133

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[bb7133]

/merge

[ti-chi-bot]

This pull request has been accepted and is ready to merge.

Commit hash: 75a772f

[bb7133]

Good job! Please update the docs @tangenta

[lysu]

do we better report duplicate errors instead of direct overwrite existing data?

for auto pk handle, it seem will report dup err:

mysql> create table t (a bigint primary key auto_increment);
Query OK, 0 rows affected (0.01 sec)

mysql> alter table t FORCE auto_increment = 20;
Query OK, 0 rows affected (0.00 sec)

mysql> insert into t values();
Query OK, 1 row affected (0.00 sec)

mysql> alter table t FORCE auto_increment = 20;
Query OK, 0 rows affected (0.01 sec)

mysql> insert into t values();
ERROR 1062 (23000): Duplicate entry '20' for key 'PRIMARY'