Issues with Centralized Allocation of Auto-Increment IDs in v6.5.0_1

Translated
1

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: v6.5.0 的中心化分配自增 ID 的问题_1

| username: GreenGuan

The concurrent execution test for the window did not reproduce the issue. It would be best to provide the script.

2
| username: WalterWj | Original post link

How does this code perform when run on MySQL? Compare the results.

3
| username: GreenGuan | Original post link

The auto-increment ID in MySQL is also not monotonically increasing based on the insertion time. What is the insertion order of the auto-increment ID generated in TiDB related to?

4
| username: Lucien-卢西恩 | Original post link

Unable to reproduce, it’s best to check if the behavior is the same on MySQL.

5
| username: GreenGuan | Original post link

The test code is as follows:

#!/usr/bin/python2.7
#-*- coding=utf8 -*-

import threading
import pymysql
import datetime
from time import ctime, sleep, time

loops = [8, 4]

dbconfig = {
    'host': 'xxx',  # vip followed by 3 TiDB compute nodes
    'port': 4100,
    'user': 'root',
    'password': 'xxx',
    'db': 'test',
    'charset': 'utf8',
    'cursorclass': pymysql.cursors.DictCursor,
}

def connTidb(sql, dbconf=dbconfig, fetchopt="all"):
    try:
        conn = pymysql.connect(**dbconf)
        cursor = conn.cursor(pymysql.cursors.DictCursor)
        conn.begin()

        cursor.execute(sql)
        if fetchopt == "all":
            res = cursor.fetchall()
        elif fetchopt == "one":
            res = cursor.fetchone()
        conn.commit()
        cursor.close()
        conn.close()
        return res
    except Exception as e:
        print(e)
        return e

def loop():
    sql = '''
    begin; insert into gyz.t1 (python_ts, tidb_ts, tso_ts) values ("{}", current_timestamp(6), @@tidb_current_ts); commit;'''.format(datetime.datetime.now())
    connTidb(sql)

def main():
    print('starting at:', ctime())
    threads = []
    nloops = range(0, 10)
    print(nloops)
    for i in nloops:
        t = threading.Thread(target=loop)
        threads.append(t)
    for i in nloops:
        threads[i].start()
    for i in nloops:
        threads[i].join()
    print('all Done at:', ctime())

if __name__ == '__main__':
    main()
6
| username: Lucien-卢西恩 | Original post link

Have you tested it on MySQL? Is the data in the id column continuous?

7
| username: GreenGuan | Original post link

The performance of MySQL and TiDB is the same, and the id may also have non-monotonic increments.

8
| username: GreenGuan | Original post link

It seems that there might be an issue with the concurrency script in Python. This problem does not occur when using Go.

9
| username: system | Original post link

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.