statistics: Remove the ineffective dirty IDs from the row count cache

[0xPoe]

What problem does this PR solve?

Issue Number: close #55803
Problem Summary:

What changed and how does it work?

As I mentioned in the issue #55803 (comment), the main problem is that UpdateByID anciently updates the modify_time even when the dirty tables have not been updated.

But as @time-and-fate mentioned, the maintenance of the dirty table follows a best-effort approach, so it would be better to delete it entirely.

Check List

Tests

• Unit test
• Integration test
• Manual test
  • statistics: Remove the ineffective dirty IDs from the row count cache #56287 (comment)
• No need to test

  • I checked and no code files have been changed.

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

修复 information_schema.tables 内存表信息不准确的问题
Fix the issue of inaccurate memory table information in `information_schema.tables`

[codecov]

Codecov Report

Attention: Patch coverage is 70.00000% with 3 lines in your changes missing coverage. Please review.

Project coverage is 73.2029%. Comparing base (ec18512) to head (f3b75ac).
Report is 3 commits behind head on master.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #56287        +/-   ##
================================================
+ Coverage   72.9851%   73.2029%   +0.2177%     
================================================
  Files          1699       1729        +30     
  Lines        469598     477589      +7991     
================================================
+ Hits         342737     349609      +6872     
- Misses       105774     106158       +384     
- Partials      21087      21822       +735     

Flag	Coverage Δ
integration	53.1788% <ø> (?)
unit	72.1618% <70.0000%> (-0.0145%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
dumpling	52.6910% <ø> (ø)
parser	∅ <ø> (∅)
br	44.9720% <ø> (+0.0196%)	⬆️

[0xPoe]

Tested in the test-infra:

mysql> SHOW STATS_META where db_name='uds500k';
+---------+-------------+----------------+---------------------+--------------+-----------+---------------------+
| Db_name | Table_name  | Partition_name | Update_time         | Modify_count | Row_count | Last_analyze_time   |
+---------+-------------+----------------+---------------------+--------------+-----------+---------------------+
| uds500k | index_Data4 |                | 2025-02-13 19:33:14 |            0 |    500000 | 2025-02-13 19:33:14 |
| uds500k | index_Data3 |                | 2025-02-13 19:33:17 |            0 |    500000 | 2025-02-13 19:33:17 |
| uds500k | index_Data1 |                | 2025-02-13 19:33:20 |            0 |    500000 | 2025-02-13 19:33:20 |
| uds500k | index_Data2 |                | 2025-02-13 19:33:23 |            0 |    500000 | 2025-02-13 19:33:23 |
| uds500k | index_Data5 |                | 2025-02-13 19:33:26 |            0 |    500000 | 2025-02-13 19:33:26 |
| uds500k | Data4       |                | 2025-02-13 19:32:58 |            0 |    500000 | 2025-02-13 19:32:58 |
| uds500k | Data5       |                | 2025-02-13 19:33:01 |            0 |    500000 | 2025-02-13 19:33:01 |
| uds500k | Data2       |                | 2025-02-13 19:33:04 |            0 |    500000 | 2025-02-13 19:33:04 |
| uds500k | Data3       |                | 2025-02-13 19:33:07 |            0 |    500000 | 2025-02-13 19:33:07 |
| uds500k | Data1       |                | 2025-02-13 19:33:10 |            0 |    500000 | 2025-02-13 19:33:10 |
+---------+-------------+----------------+---------------------+--------------+-----------+---------------------+
10 rows in set (0.39 sec)

mysql> select table_name, avg_row_length, max_data_length, data_length, table_rows from information_schema.tables where table_schema = 'uds500k';
+-------------+----------------+-----------------+-------------+------------+
| table_name  | avg_row_length | max_data_length | data_length | table_rows |
+-------------+----------------+-----------------+-------------+------------+
| index_Data4 |            135 |               0 |    67991573 |     500000 |
| index_Data3 |            136 |               0 |    68001335 |     500000 |
| index_Data1 |            136 |               0 |    68002490 |     500000 |
| index_Data2 |            136 |               0 |    68003939 |     500000 |
| index_Data5 |            136 |               0 |    68004062 |     500000 |
| Data4       |             24 |               0 |    12000000 |     500000 |
| Data5       |             24 |               0 |    12000000 |     500000 |
| Data2       |             24 |               0 |    12000000 |     500000 |
| Data3       |             24 |               0 |    12000000 |     500000 |
| Data1       |             24 |               0 |    12000000 |     500000 |
+-------------+----------------+-----------------+-------------+------------+
10 rows in set (0.08 sec)

[0xPoe]

🔢 Self-check (PR reviewed by myself and ready for feedback.)

[0xPoe]

Tested locally:

tiup playground v8.5.1 --db.host 127.0.0.1 --without-monitor --tiflash 0 --db 2

On tidb1:

use test;
create table t1(a int);create table t3(a int);create table t2(a int);
insert into t1 value(1);
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t1 select * from t1;
insert into t2 select * from t1;
insert into t3 select * from t1;

Wait until on tidb1:

show stats_meta;

mysql> show stats_meta;
+---------+------------+----------------+---------------------+--------------+-----------+-------------------+
| Db_name | Table_name | Partition_name | Update_time         | Modify_count | Row_count | Last_analyze_time |
+---------+------------+----------------+---------------------+--------------+-----------+-------------------+
| test    | t1         |                | 2025-02-26 14:39:12 |        16384 |     16384 | NULL              |
| test    | t3         |                | 2025-02-26 14:39:12 |        16384 |     16384 | NULL              |
| test    | t2         |                | 2025-02-26 14:39:12 |        16384 |     16384 | NULL              |
+---------+------------+----------------+---------------------+--------------+-----------+-------------------+
3 rows in set (0.02 sec)

Then, on tidb2:

mysql> select table_name, avg_row_length, max_data_length, data_length, table_rows from information_schema.tables where table_schema = 'test';
+------------+----------------+-----------------+-------------+------------+
| table_name | avg_row_length | max_data_length | data_length | table_rows |
+------------+----------------+-----------------+-------------+------------+
| t1         |              8 |               0 |      131072 |      16384 |
| t3         |              8 |               0 |      131072 |      16384 |
| t2         |              8 |               0 |      131072 |      16384 |
+------------+----------------+-----------------+-------------+------------+
3 rows in set (0.02 sec)

[time-and-fate]

Please add the reason.

[time-and-fate]

Others LGTM.

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: qw4990, tangenta, time-and-fate

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [qw4990,tangenta,time-and-fate]
• pkg/statistics/OWNERS [qw4990,time-and-fate]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[0xPoe]

/test all

[0xPoe]

/retest

[ti-chi-bot]

In response to a cherrypick label: new pull request created to branch release-7.1: #59854.
But this PR has conflicts, please resolve them!

[ti-chi-bot]

In response to a cherrypick label: new pull request created to branch release-8.1: #59855.
But this PR has conflicts, please resolve them!

[ti-chi-bot]

In response to a cherrypick label: new pull request created to branch release-8.5: #59862.