MySQL DeadLock故障排查全过程记录

时间:2019-02-27 09:45:43          访问次数:0

【作者】

刘博:携程技术保障中心数据库高级经理,主要关注Sql server和Mysql的运维和故障处理。

【环境】

版本号:5.6.21

隔离级别:REPEATABLE READ

【问题描述】

接到监控报警,有一个线上的应用DeadLock报错,每15分钟会准时出现,报错统计如下图:

登录Mysql服务器查看日志:

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

mysql> show engine innodb status\G

 

*** (1) TRANSACTION:

 

TRANSACTION 102973, ACTIVE 11 sec starting index read

 

mysql tables in use 3, locked 3

 

LOCK WAIT 4 lock struct(s), heap size 1136, 3 row lock(s)

 

MySQL thread id 6, OS thread handle 140024996574976, query id 83 localhost us updating

 

UPDATE TestTable

 

SET column1 = 1,

 

Column2 = sysdate(),

 

Column3= '026'

 

Column4 = 0

 

AND column5 = 485

 

AND column6 = 'SEK'

 

*** (1) WAITING FOR THIS LOCK TO BE GRANTED:

 

RECORD LOCKS space id 417 page no 1493 n bits 1000 index idx_column6 of table test.TestTable trx id 102973 lock_mode X waiting

 

Record lock, heap no 859 PHYSICAL RECORD: n_fields 2; compact format; info bits 0

 

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007e1452; asc ~ R;;

 

*** (2) TRANSACTION:

 

TRANSACTION 102972, ACTIVE 26 sec starting index read

 

mysql tables in use 3, locked 3

 

219 lock struct(s), heap size 24784, 2906 row lock(s), undo log entries 7

 

MySQL thread id 5, OS thread handle 140024996841216, query id 84 localhost us updating

 

UPDATE TestTable

 

SET Column1 = 1,

 

Column2 = sysdate(),

 

Column3 = '026'

 

Column4 = 0

 

AND Column5 = 485

 

AND Column6 = 'SEK'

 

*** (2) HOLDS THE LOCK(S):

 

RECORD LOCKS space id 417 page no 1493 n bits 1000 index idx_Column6 of table test.TestTable trx id 102972 lock_mode X

 

Record lock, heap no 1 PHYSICAL RECORD: n_fields 1; compact format; info bits 0

 

0: len 8; hex 73757072656d756d; asc supremum;;

 

 

Record lock, heap no 859 PHYSICAL RECORD: n_fields 2; compact format; info bits 0

 

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007e1452; asc ~ R;;

?

1

2

3

4

5

6

7

8

9

*** (2) WAITING FOR THIS LOCK TO BE GRANTED:

 

RECORD LOCKS space id 601 page no 89642 n bits 1000 index idx_column6 of table test.TestTable trx id 32231892482 lock_mode X locks rec but not gap waiting

 

Record lock, heap no 38 PHYSICAL RECORD: n_fields 2; compact format; info bits 0

 

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007eea14; asc ~ ;;

大致一看,更新同一索引的同一行,应该是一个Block,报TimeOut的错才对,怎么会报DeadLock?

【初步分析】

先分析下(2) TRANSACTION,TRANSACTION 32231892482。

等待的锁信息为:

?

1

2

3

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007eea14; asc

持有的锁信息为:

?

1

2

3

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007eeac4; asc

再先分析下(1) TRANSACTION,TRANSACTION 32231892617。

等待的锁信息为:

?

1

2

3

0: len 3; hex 53454b; asc SEK;;

 

1: len 8; hex 80000000007eeac4; asc

于是可以画出的死锁表,两个资源相互依赖,造成死锁:

 

TRANSACTION Hold Wait
32231892617 53454b\80000000007eea14 53454b\80000000007eeac4
32231892482 53454b\80000000007eeac4 53454b\80000000007eea14

 

让我们再看一下explain结果:

mysql>desc UPDATE TestTable SET Column1=1, Column2 = sysdate(),Column3 = '025' Column4 = 0 AND Column5 = 477 AND Column6 = 'SEK' \G;

*************************** 1. row ***************************

id: 1

select_type: UPDATE

table: TestTable

partitions: NULL

type: index_merge

possible_keys: column5_index,idx_column5_column6_Column1,idxColumn6

key: column5_index,idxColumn6

key_len: 8,9

ref: NULL

rows: 7

filtered: 100.00

Extra: Using intersect(column5_index,idxColumn6); Using where

可以看到 EXTRA 列:

?

1

Using intersect(column5_index,idxColumn6)

从5.1开始,引入了 index merge 优化技术,对同一个表可以使用多个索引分别进行条件扫描。

相关文档:http://dev.mysql.com/doc/refman/5.7/en/index-merge-optimization.html

The Index Merge method is used to retrieve rows with several range scans and to merge their results into one. The merge can produce unions, intersections, or unions-of-intersections of its underlying scans. This access method merges index scans from a single table; it does not merge scans across multiple tables.

【模拟与验证】

根据以上初步分析,猜测应该就是intersect造成的,于是在测试环境模拟验证,开启2个session模拟死锁:

 

时间序列 Session1 Session2
1 Begin;  
2 UPDATE TestTable SET Column2 = sysdate() Column4 = 0 AND Column5 = 47 AND Column6 = 'SEK 
执行成功,影响7行
 
3   Begin;
4   UPDATE TestTable SET Column2 = sysdate(),Column4 = 0 AND Column5 = 485 AND Column6 = 'SEK';
被Blocking
5 UPDATE TestTable SET Column2 = sysdate(),Column4 = 0 AND Column5 = 485 AND Column6 = 'SEK';
执行成功
ERROR 1213 (40001): Deadlock found when trying to get lock; try restarting transaction

 

依据以上信息可以发现Session2虽然被Block了,但也获取了一些Session1在时间序列5时所需资源的X锁,可以再开启一个查询select count(Column5) from TestTable where Column5 = 485,设置SET TRANSACTION ISOLATION LEVEL SERIALIZABLE,去查询Column5 = 485的行,观察锁等待的信息:
mysql> SELECT r.trx_id waiting_trx_id, r.trx_mysql_thread_id waiting_thread, r.trx_query waiting_query, b.trx_id

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

mysql> SELECT r.trx_id waiting_trx_id, r.trx_mysql_thread_id waiting_thread, r.trx_query waiting_query, b.trx_id blocking_trx_id, b.trx_mysql_thread_id blocking_thread, b.trx_query blocking_query FROM information_schema.innodb_lock_waits w INNER JOIN information_schema.innodb_trx b ON b.trx_id = w.blocking_trx_id INNER JOIN information_schema.innodb_trx r ON r.trx_id = w.requesting_trx_id \G;

 

*************************** 1. row ***************************

 

waiting_trx_id: 103006

 

waiting_thread: 36

 

waiting_query: UPDATE TestTable SET Column1 = 1, Column2 = sysdate(), Column3 = '026' Column4 = 0 AND Column5 = 485 AND Column6 = 'SEK'

 

blocking_trx_id: 103003

 

blocking_thread: 37

 

blocking_query: NULL

 

*************************** 2. row ***************************

 

waiting_trx_id: 421500433538672

 

waiting_thread: 39

 

waiting_query: select count(Column5) from TestTable where Column5 = 485

 

blocking_trx_id: 103006

 

blocking_thread: 36

 

blocking_query: UPDATE TestTable SET Column1 = 1, Column2 = sysdate(), Column3 = '026' Column4 = 0 AND Column5 = 485 AND Column6 = 'SEK'

 

2 rows in set, 1 warning (0.00 sec)

 

mysql> select * from information_schema.innodb_lock_waits \G;

 

*************************** 1. row ***************************

 

requesting_trx_id: 103006

 

requested_lock_id: 103006:417:1493:859

 

blocking_trx_id: 103003

 

blocking_lock_id: 103003:417:1493:859

 

*************************** 2. row ***************************

 

requesting_trx_id: 421500433538672

 

requested_lock_id: 421500433538672:417:749:2

 

blocking_trx_id: 103006

 

blocking_lock_id: 103006:417:749:2

 

2 rows in set, 1 warning (0.00 sec)

?

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

mysql> select * from INNODB_LOCKS \G;

 

*************************** 1. row ***************************

 

lock_id: 103006:417:1493:859

 

lock_trx_id: 103006

 

lock_mode: X

 

lock_type: RECORD

 

lock_table: test.TestTable

 

lock_index: idxColumn6

 

lock_space: 417

 

lock_page: 1493

 

lock_rec: 859

 

lock_data: 'SEK', 8262738

 

*************************** 2. row ***************************

 

lock_id: 103003:417:1493:859

 

lock_trx_id: 103003

 

lock_mode: X

 

lock_type: RECORD

 

lock_table:test.TestTable

 

lock_index: idxColumn6

 

lock_space: 417

 

lock_page: 1493

 

lock_rec: 859

 

lock_data: 'SEK', 8262738

 

*************************** 3. row ***************************

 

lock_id: 421500433538672:417:749:2

 

lock_trx_id: 421500433538672

 

lock_mode: S

 

lock_type: RECORD

 

lock_table: test.TestTable

 

lock_index: column5_index

 

lock_space: 417

 

lock_page: 749

 

lock_rec: 2

 

lock_data: 485, 8317620

 

*************************** 4. row ***************************

 

lock_id: 103006:417:749:2

 

lock_trx_id: 103006

 

lock_mode: X

 

lock_type: RECORD

 

lock_table: test.TestTable

 

lock_index: column5_index

 

lock_space: 417

 

lock_page: 749

 

lock_rec: 2

 

lock_data: 485, 8317620

 

4 rows in set, 1 warning (0.00 sec)

可以看到Session2,trx_id 103006阻塞了trx_id 421500433538672,而trx_id 421500433538672 requested_lock也正好是lock_data: 485, 8317620。由此可见Session2虽然别block了,但是还是获取到了Index column5_index相关的锁。被Block是因为intersect的原因,还需要idxColumn6的锁,至此思路已经清晰,对整个分配锁的信息简化一下,如下表格(请求到的锁用青色表示,需获取但未获取到的锁用红色表示):

 

时间点 Session1 Session2
1 477 SEK  
2   485 SEK
3 485 SEK 死锁发生

 

可以看到485 SEK这两个资源形成了一个环状,最终发生死锁。

【解决方法】

  • 最佳的方法是添加column5和Column6的联合索引。
  • 我们环境当时的情况发现Column6的筛选度非常低,就删除了Column6的索引。
    10:55左右删除索引后,报错没有再发生:

总结

以上就是这篇文章的全部内容了,希望本文的内容对大家的学习或者工作具有一定的参考学习价值,如果有疑问大家可以留言交流,谢谢大家对脚本之家的支持。