lightning max-error record tripled when backend is tidb

[fubinzh]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

1. Do lightning import for data with 2 type errors (e.g. type error), and check max-error records in lightning_task_info.type_error_v1

Test data:

[root@centos76_vm type_error]# tail max_error.t2.1.csv
21348226,"lGztBz","2021-11-17 00:00:00"
89250625,"ltDGh","2021-11-17 00:00:00"
9207504,"iEMNwuP","2021-11-17 00:00:00"
1,"b1","2021-11-14 00:00:00"
2,"b2","2021-11-14 00:00:00"
3,"b3","2021-11-14 00:00:00"
4,"b4","2021-11-14 00:00:00"
"xxx","b5","2021-11-15 00:00:00"
5,"b6","0"

[root@centos76_vm type_error]# cat max_error.t2-schema.sql
/*!40101 SET NAMES binary*/;
CREATE TABLE `t2` (
  `a` int(11) NOT NULL,
  `b` varchar(10) DEFAULT NULL,
  `c` datetime DEFAULT NULL,
  PRIMARY KEY (`a`) /*T![clustered_index] CLUSTERED */,
  UNIQUE KEY `uk_b` (`b`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin;

lightning config is as below

[root@centos76_vm type_error]# grep -v ^# lightning.toml
[lightning]
level = "info"
file = "tidb-lightning.log"
max-error = 10

[checkpoint]
driver = "mysql"

[tikv-importer]
backend = "tidb"
sorted-kv-dir = "/home/sorted-kv-dir"

[mydumper]
data-source-dir = "s3://nfs/tools_test_data/max_error/type_error?access-key=minioadmin&secret-access-key=minioadmin&endpoint=http%3a%2f%2fminio.pingcap.net%3a9000&force-path-style=true"
no-schema = false

filter = ['*.*', 'mysql.*', '!sys.*', '!INFORMATION_SCHEMA.*', '!PERFORMANCE_SCHEMA.*', '!METRICS_SCHEMA.*', '!INSPECTION_SCHEMA.*']

[mydumper.csv]
separator = ','
delimiter = '"'
header = true


[tidb]
host = "172.16.6.217"
port = 4000
user = "root"
status-port = 10080
sql-mode = "STRICT_TRANS_TABLES"

2. What did you expect to see? (Required)

lightning_task_info.type_error_v1 should record 2 rows with type error

3. What did you see instead (Required)

lightning_task_info.type_error_v1 recorded 6 rows with type error, the records are tripled.
This issue is NOT seem when backend is local.

mysql> select * from lightning_task_info.type_error_v1;
+---------------------+----------------------------+------------------+--------------------+--------+-------------------------------------------------------------------+------------------------------------+
| task_id             | create_time                | table_name       | path               | offset | error                                                             | row_data                           |
+---------------------+----------------------------+------------------+--------------------+--------+-------------------------------------------------------------------+------------------------------------+
| 1637567621075919308 | 2021-11-22 15:53:42.370553 | `max_error`.`t2` | max_error.t2.1.csv |   3803 | Error 1366: Incorrect int value: 'xxx' for column 'a' at row 1    | ('xxx','b5','2021-11-15 00:00:00') |
| 1637567621075919308 | 2021-11-22 15:53:42.379626 | `max_error`.`t2` | max_error.t2.1.csv |   3836 | Error 1292: Incorrect datetime value: '0' for column 'c' at row 1 | ('5','b6','0')                     |
| 1637567621075919308 | 2021-11-22 15:53:42.507533 | `max_error`.`t2` | max_error.t2.1.csv |   3803 | Error 1366: Incorrect int value: 'xxx' for column 'a' at row 1    | ('xxx','b5','2021-11-15 00:00:00') |
| 1637567621075919308 | 2021-11-22 15:53:42.512061 | `max_error`.`t2` | max_error.t2.1.csv |   3836 | Error 1292: Incorrect datetime value: '0' for column 'c' at row 1 | ('5','b6','0')                     |
| 1637567621075919308 | 2021-11-22 15:53:42.640734 | `max_error`.`t2` | max_error.t2.1.csv |   3803 | Error 1366: Incorrect int value: 'xxx' for column 'a' at row 1    | ('xxx','b5','2021-11-15 00:00:00') |
| 1637567621075919308 | 2021-11-22 15:53:42.645454 | `max_error`.`t2` | max_error.t2.1.csv |   3836 | Error 1292: Incorrect datetime value: '0' for column 'c' at row 1 | ('5','b6','0')                     |
+---------------------+----------------------------+------------------+--------------------+--------+-------------------------------------------------------------------+------------------------------------+
6 rows in set (0.00 sec)

4. What is your TiDB version? (Required)

[root@centos76_vm type_error]# tiup tidb-lightning:v5.3.0 -V
Starting component tidb-lightning: /root/.tiup/components/tidb-lightning/v5.3.0/tidb-lightning -V
Release Version: v5.3.0
Git Commit Hash: 79e237d
Git Branch: heads/refs/tags/v5.3.0
Go Version: go1.16.4
UTC Build Time: 2021-11-18 13:06:05
Race Enabled: false

[github-actions]

Please check whether the issue should be labeled with 'affects-x.y' or 'fixes-x.y.z', and then remove 'needs-more-info' label.