如何解决使用AWS Glue作业的MySQL RDS中的死锁问题
我已经在AWS UI上使用pyspark编写了Glue作业。该作业从json中提取列名称和值,然后写入MySQL表。 直到昨天,一切都很好。现在,我在AWS Glue控制台上收到错误消息:
调用o984.pyWriteDynamicFrame时发生错误。作业由于阶段失败而中止:阶段371.0中的任务71失败了4次,最近一次失败:阶段371.0中的任务71.3丢失(TID 25999,ip-10-0-1-5.eu-west-1.compute.internal,执行者7):UnknownReason
在cloudwatch日志中,我可以看到: java.sql.BatchUpdateException:尝试获取锁时发现了死锁;尝试重新启动事务
当我在MySQL RDS实例上运行命令时:
**show engine innodb status;**
Output:
LATEST DETECTED DEADLOCK
------------------------
2020-08-11 10:17:57 0x2aca59649700TOO DEEP OR LONG SEARCH IN THE LOCK TABLE WAITS-FOR GRAPH,WE WILL ROLL BACK FOLLOWING TRANSACTION
*** TRANSACTION:
TRANSACTION 71219,ACTIVE 0 sec inserting
mysql tables in use 1,locked 1
6 lock struct(s),heap size 1136,4 row lock(s),undo log entries 75
MySQL thread id 1771,OS thread handle 47048571524864,query id 38693 10.0.1.175 admin update
INSERT INTO etl_adp_engine_audit_items (`objectid`,`formula`,`date_end`,`decline_reasonid`,`object_name`,`audit_levelid`,`date_start`,`object_no`,`is_exit`,`outcomeid`,`parent_objectid`,`resultid`,`object_levelid`,`application_id`,`object_type`,`etl_adp_audit_id`,`id`) VALUES ('ab072fbc-88ce-4516-b89a-e44a5bb1f19d','ADP_Engine_Logic_Action_Variables.CreateVariableInteger(Convert.ToString(Functions.ParseEquation(Me,\'SC01\')),System.Int32.Parse(Functions.ParseEquation(Me,\'264\')))','2019-03-11 12:44:49.443000',NULL,'5.001 - Create Variable SC01',1,253,'026f2cec-253f-4ac9-a4b9-3195de2a6a11',2,'1552308273','Action',20317,3365)
*** WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 443 page no 10 n bits 240 index adp_audit_items_index of table `culoansportaldb`.`etl_adp_engine_audit_items` trx id 71219 lock_mode X locks gap before rec insert intention waiting
Record lock,heap no 43 PHYSICAL RECORD: n_fields 5; compact format; info bits 0
0: len 27; hex 352e3031202d2048617665457870657269616e203d202759455327; asc 5.01 - HaveExperian = 'YES';;
1: len 4; hex 80000108; asc ;;
2: len 4; hex 52756c65; asc Rule;;
3: len 8; hex 8000000000100ea6; asc ;;
4: len 8; hex 800000000000965b; asc [;;
*** WE ROLL BACK TRANSACTION (2)
------------
TRANSACTIONS
------------
Trx id counter 71241
Purge done for trx's n:o < 71238 undo n:o < 0 state: running but idle
History list length 10
LIST OF TRANSACTIONS FOR EACH SESSION:
---TRANSACTION 328521077494880,not started
0 lock struct(s),0 row lock(s)
---TRANSACTION 328521077494016,0 row lock(s)
---TRANSACTION 328521077493152,0 row lock(s)
---TRANSACTION 328521077492288,0 row lock(s)
---TRANSACTION 328521077491424,0 row lock(s)
谁能告诉我如何在Glue中解决问题,或者必须在MySQL上解决问题?
我已经研究过AWS Glue,到目前为止,还没有像自动提交或事务回滚这样的支持。
任何帮助将不胜感激!
解决方法
我相信这是胶水问题,但是您可以在MySQL中转过身来。
您可以将参数innodb_lock_wait_timeout增大到最大值,以查看它是否对您有用。
并且,如果可能的话,对粘合执行器的数量设置较低的限制。
版权声明:本文内容由互联网用户自发贡献,该文观点与技术仅代表作者本人。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容, 请发送邮件至 dio@foxmail.com 举报,一经查实,本站将立刻删除。