故障解析丨Clone节点导致主从故障

牢记使命不忘初心

1.背景概述

在一次主从复制架构中，由于主节点binlog损坏，导致从节点无法正常同步数据，只能重做从节点；因此使用MySQL 8.0.17开始提供的clone技术进行恢复，恢复后的2天都发生了主从报错数据冲突。
通过解析binlog发现，同一时刻主从节点都在执行同一条语句，因此询问业务是否在主从节点都执行了定时任务，业务回复定时任务只在主节点执行。
最后排查发现，克隆后的从节点的定时任务也会是开启的状态，因此同一时刻，主从节点同时执行定时任务，导致主从报错，最终将从节点的定时任务关闭后解决此问题。
2.问题复现

本次测试基于 GreatSQL 8.0.32-24

1. greatsql> SELECT VERSION();
2. +-----------+
3. | VERSION() |
4. +-----------+
5. | 8.0.32-24 |
6. +-----------+
7. 1 row in set (0.00 sec)

复制代码

1.搭建一套主从架构

略
2.创建event

1. greatsql> create database test;
2. greatsql> use test;
3. greatsql> CREATE TABLE `test` (
4. `id` int(11) unsigned NOT NULL AUTO_INCREMENT COMMENT 'ID',
5. `now` datetime DEFAULT NULL COMMENT '时间',
6. PRIMARY KEY (`id`)
7. );
9. greatsql> CREATE EVENT event_test
10. ON SCHEDULE EVERY 1 MINUTE
11. ON COMPLETION PRESERVE
12. ENABLE
13. COMMENT '每隔1分钟向test表插入记录'
14. DO INSERT INTO test VALUES(NULL, now());

复制代码

3.查看event状态

主节点，默认情况下event状态为 ENABLED

1. greatsql> show events;
2. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+----------------------+----------------------+--------------------+
3. | Db  | Name    | Definer | Time zone | Type    | Execute at | Interval value | Interval field | Starts        | Ends | Status  | Originator | character_set_client | collation_connection | Database Collation |
4. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+----------------------+----------------------+--------------------+
5. | test | event_test | root@%  | SYSTEM   | RECURRING | NULL    | 1        | MINUTE     | 2023-10-12 17:11:14 | NULL | ENABLED |      1 | utf8mb4        | utf8mb4_unicode_ci  | utf8mb4_unicode_ci |
6. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+----------------------+----------------------+--------------------+
7. 1 row in set (0.00 sec)

复制代码

从节点，默认情况下event状态为 SLAVESIDE_DISABLED

1. greatsql> show events;
2. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+--------------------+------------+----------------------+----------------------+--------------------+
3. | Db  | Name    | Definer | Time zone | Type    | Execute at | Interval value | Interval field | Starts        | Ends | Status       | Originator | character_set_client | collation_connection | Database Collation |
4. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+--------------------+------------+----------------------+----------------------+--------------------+
5. | test | event_test | root@%  | SYSTEM   | RECURRING | NULL    | 1        | MINUTE     | 2023-10-12 17:11:14 | NULL | SLAVESIDE_DISABLED |      1 | utf8mb4        | utf8mb4_unicode_ci  | utf8mb4_unicode_ci |
6. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+--------------------+------------+----------------------+----------------------+--------------------+
7. 1 row in set (0.00 sec)

复制代码

4.查看数据

1. greatsql> select * from test.test;
2. +----+---------------------+
3. | id | now                 |
4. +----+---------------------+
5. |  1 | 2023-08-08 16:00:39 |
6. |  2 | 2023-08-08 16:01:39 |
7. |  3 | 2023-08-08 16:02:39 |
8. +----+---------------------+
9. 3 rows in set (0.00 sec)

复制代码

5.从节点进行克隆

# 安装克隆插件，主从节点都需要

1. greatsql> install plugin clone soname 'mysql_clone.so';

复制代码

# 从节点进行clone

1. greatsql> set global clone_valid_donor_list='172.17.137.162:6001';
3. greatsql> clone instance from root@'172.17.137.162':6001 identified by 'greatsql';

复制代码

6.重新建立主从复制

1. greatsql> change master to master_user='root',master_password='greatsql',master_host='172.17.137.162',master_port=6001,master_auto_position=1;
2. Query OK, 0 rows affected, 7 warnings (0.04 sec)
4. greatsql> start slave;
5. Query OK, 0 rows affected, 1 warning (0.04 sec)

复制代码

7.查看主从状态

1. greatsql> show slave status\G
2. *************************** 1. row ***************************
3.                Slave_IO_State: Waiting for source to send event
4.                   Master_Host: 172.17.137.162
5.                   Master_User: root
6.                   Master_Port: 6001
7.                 Connect_Retry: 60
8.               Master_Log_File: binlog.000001
9.           Read_Master_Log_Pos: 2959
10.                Relay_Log_File: relaylog.000002
11.                 Relay_Log_Pos: 395
12.         Relay_Master_Log_File: binlog.000001
13.              Slave_IO_Running: Yes
14.             Slave_SQL_Running: No
15.               Replicate_Do_DB:
16.           Replicate_Ignore_DB:
17.            Replicate_Do_Table:
18.        Replicate_Ignore_Table:
19.       Replicate_Wild_Do_Table:
20.   Replicate_Wild_Ignore_Table:
21.                    Last_Errno: 1062
22.                    Last_Error: Coordinator stopped because there were error(s) in the worker(s). The most recent failure being: Worker 1 failed executing transaction 'e8bf88f9-2acd-11ee-a98a-00163e605c74:8' at master log binlog.000001, end_log_pos 2606. See error log and/or performance_schema.replication_applier_status_by_worker table for more details about this failure or others, if any.
23.                  Skip_Counter: 0
24.           Exec_Master_Log_Pos: 2307
25.               Relay_Log_Space: 1242
26.               Until_Condition: None
27.                Until_Log_File:
28. greatsql> select * from performance_schema.replication_applier_status_by_worker limit 1\G
29. *************************** 1. row ***************************
30.                                            CHANNEL_NAME:
31.                                               WORKER_ID: 1
32.                                               THREAD_ID: NULL
33.                                           SERVICE_STATE: OFF
34.                                       LAST_ERROR_NUMBER: 1062
35.                                      LAST_ERROR_MESSAGE: Worker 1 failed executing transaction 'e8bf88f9-2acd-11ee-a98a-00163e605c74:8' at master log binlog.000001, end_log_pos 2606; Could not execute Write_rows event on table test.test; Duplicate entry '5' for key 'test.PRIMARY', Error_code: 1062; handler error HA_ERR_FOUND_DUPP_KEY; the event's master log FIRST, end_log_pos 2606
36.                                    LAST_ERROR_TIMESTAMP: 2023-08-08 16:03:39.033240
37.                                LAST_APPLIED_TRANSACTION:
38.      LAST_APPLIED_TRANSACTION_ORIGINAL_COMMIT_TIMESTAMP: 0000-00-00 00:00:00.000000
39.     LAST_APPLIED_TRANSACTION_IMMEDIATE_COMMIT_TIMESTAMP: 0000-00-00 00:00:00.000000
40.          LAST_APPLIED_TRANSACTION_START_APPLY_TIMESTAMP: 0000-00-00 00:00:00.000000
41.            LAST_APPLIED_TRANSACTION_END_APPLY_TIMESTAMP: 0000-00-00 00:00:00.000000
42.                                    APPLYING_TRANSACTION: e8bf88f9-2acd-11ee-a98a-00163e605c74:8
43.          APPLYING_TRANSACTION_ORIGINAL_COMMIT_TIMESTAMP: 2023-08-08 16:02:45.795753
44.         APPLYING_TRANSACTION_IMMEDIATE_COMMIT_TIMESTAMP: 2023-08-08 16:02:45.795753
45.              APPLYING_TRANSACTION_START_APPLY_TIMESTAMP: 2023-08-08 16:03:39.032510
46.                  LAST_APPLIED_TRANSACTION_RETRIES_COUNT: 0
47.    LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_NUMBER: 0
48.   LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_MESSAGE:
49. LAST_APPLIED_TRANSACTION_LAST_TRANSIENT_ERROR_TIMESTAMP: 0000-00-00 00:00:00.000000
50.                      APPLYING_TRANSACTION_RETRIES_COUNT: 0
51.        APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_NUMBER: 0
52.       APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_MESSAGE:
53.     APPLYING_TRANSACTION_LAST_TRANSIENT_ERROR_TIMESTAMP: 0000-00-00 00:00:00.000000
54. 1 row in set (0.00 sec）

复制代码

可以看到从节点报错发生了主键冲突。
8.查看从节点定时任务状态

当前从节点定时任务状态为 ENABLED

1. greatsql> show events;
2. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+----------------------+----------------------+--------------------+
3. | Db  | Name    | Definer | Time zone | Type    | Execute at | Interval value | Interval field | Starts        | Ends | Status  | Originator | character_set_client | collation_connection | Database Collation |
4. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+----------------------+----------------------+--------------------+
5. | test | event_test | root@%  | SYSTEM   | RECURRING | NULL    | 1        | MINUTE     | 2023-08-08 15:58:45 | NULL | ENABLED |      1 | utf8mb4        | utf8mb4_unicode_ci  | utf8mb4_unicode_ci |
6. +------+------------+---------+-----------+-----------+------------+----------------+----------------+---------------------+------+---------+------------+---------------------+----------------------+--------------------+
7. 1 row in set (0.00 sec)

复制代码

可以看到由于从节点的定时任务也执行了，从节点写入数据，导致主键冲突。
9.故障解决

1. greatsql> alter event event_test DISABLE;
2. Query OK, 0 rows affected (0.01 sec)

复制代码

关闭从节点的定时任务event，然后跳过主键冲突的报错，最后重新启动主从复制。
3.总结

1.如果主库有定时任务，通过clone的方式搭建从库，在从库恢复之后需要关闭定时任务，避免主从同时执行定时任务导致主从故障。
2.克隆时，如果捐赠节点有主从复制信息，则克隆后的接收节点也会克隆此复制信息，并在克隆完成自动重启实例后，自动启动复制；避免此问题可以在接收节点的配置文件中增加 skip-slave-start，避免节点重启后自动启动复制。

Enjoy GreatSQL
来源:https://www.cnblogs.com/greatsql/p/17786467.html
免责声明：由于采集信息均来自互联网，如果侵犯了您的权益，请联系我们【E-Mail:cb@itdo.tech】 我们会及时删除侵权内容，谢谢合作！