mysql 5.6 read-committed隔离级别下并发插入唯一索引导致死锁一例

Posted 2020-08-19 太白的技术博客

今天，某个环境又发生了死锁，如下：

*** (1) TRANSACTION:
TRANSACTION 735307073, ACTIVE 0 sec inserting
mysql tables in use 1, locked 1
LOCK WAIT 6 lock struct(s), heap size 1184, 3 row lock(s), undo log entries 1
MySQL thread id 2754, OS thread handle 0x7f29cd89a700, query id 751744317 127.0.0.1 osm update
insert into tb_trd_secu_command(
machine_date, company_no, product_id,
product_code, product_name, unit_id, unit_code,
unit_name, asset_account, stock_account, command_operator,
command_executor, command_date, command_time, command_id,
batch_no, exchange_no, asset_type, stock_type,
stock_code, stock_name, currency_type, command_dir,
limit_price, command_qty, command_amt, order_qty,
cancel_qty, command_status, execution_status, execution_amt,
execution_qty, frozen_amt)
value(
v_machine_date, v_company_no, v_product_id,
v_product_code, v_product_name, v_unit_id, v_unit_code,
v_unit_name, v_asset_account, v_stock_account, v_command_operator,
v_command_executor, v_command_date, v_command_time, v_command_id,
v_batch_no, v_exchange_no, v_asset_type, v_stock_type,
*** (1) WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 10538 page no 4 n bits 296 index `idx_tb_trd_secu_command_1` of table `db_trd`.`tb_trd_secu_command` trx id 735307073 lock_mode X locks gap before rec insert intention waiting
*** (2) TRANSACTION:
TRANSACTION 735307149, ACTIVE 0 sec inserting, thread declared inside InnoDB 1
mysql tables in use 1, locked 1
4 lock struct(s), heap size 1184, 2 row lock(s), undo log entries 1
MySQL thread id 2744, OS thread handle 0x7f29cd8db700, query id 751750715 127.0.0.1 osm update
insert into tb_trd_secu_command(
machine_date, company_no, product_id,
product_code, product_name, unit_id, unit_code,
unit_name, asset_account, stock_account, command_operator,
command_executor, command_date, command_time, command_id,
batch_no, exchange_no, asset_type, stock_type,
stock_code, stock_name, currency_type, command_dir,
limit_price, command_qty, command_amt, order_qty,
cancel_qty, command_status, execution_status, execution_amt,
execution_qty, frozen_amt)
value(
v_machine_date, v_company_no, v_product_id,
v_product_code, v_product_name, v_unit_id, v_unit_code,
v_unit_name, v_asset_account, v_stock_account, v_command_operator,
v_command_executor, v_command_date, v_command_time, v_command_id,
v_batch_no, v_exchange_no, v_asset_type, v_stock_type,
*** (2) HOLDS THE LOCK(S):
RECORD LOCKS space id 10538 page no 4 n bits 296 index `idx_tb_trd_secu_command_1` of table `db_trd`.`tb_trd_secu_command` trx id 735307149 lock mode S locks gap before rec
*** (2) WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 10538 page no 4 n bits 296 index `idx_tb_trd_secu_command_1` of table `db_trd`.`tb_trd_secu_command` trx id 735307149 lock_mode X locks gap before rec insert intention waiting
*** WE ROLL BACK TRANSACTION (2)

其中idx_tb_trd_secu_command_1是唯一索引。

看了下参数，当前隔离级别是read-committed，不过参数innodb_locks_unsafe_for_binlog是默认值OFF。

又仔细看了下官方文档， innodb_locks_unsafe_for_binlog和read-committed一方面是一样的，启用后，可以使得InnoDB gap锁最小化，但是在两种场景(外键约束和唯一索引)中，仍然不可避免的存在gap锁。

不管事务管理级别如何设置，只要存在唯一性约束并且可能insert/delete/update同一key值的记录，死锁都将无法100%的避免，只能是概率降低。