users table has two unique keys defined ('id' and 'login' columns, two first columns in value list) and they both have value passed in the value list.
Has the definition of users table changed? I wonder if I'm using wrong schema definition in test cluster...
btw. I'm now growing the size of user_best_times table in the testing to see if problem is triggered by row count somehow. I'm now at 30M rows - no issues so far.
These NULL key warnings don't make sense:
130210 13:21:06 [Note] WSREP: NULL key skipped: INSERT IGNORE INTO users VALUES ('14214700' ,'iph9027E4342B 35','<skipped for security reasons> ','','' ,NULL,NULL, '','0', ' ','','' ,'','', NULL,'2013- 02-10 13:21:06' ,'1965' ,'0','0' ,NULL,NULL, NULL,NULL, '0',NULL, NULL,NULL, NULL,'2013- 02-10 13:21:06' ,NULL,NULL, NULL,'0. 00',NULL, '2006-01- 01',NULL, '0',NULL, '0','0' ,NULL,NULL, NULL,'0' ,NULL,' 0.0000' ,'0.00' ,'2012- 08-10 13:21:06' ,NULL,' 2','1', '2',NULL, NULL,'0000- 00-00 00:00:00' ,NULL,NULL, NULL,NULL, NULL,NULL, '0000-01- 01 00:00:00' ,NULL,NULL, '1')
users table has two unique keys defined ('id' and 'login' columns, two first columns in value list) and they both have value passed in the value list.
Has the definition of users table changed? I wonder if I'm using wrong schema definition in test cluster...
btw. I'm now growing the size of user_best_times table in the testing to see if problem is triggered by row count somehow. I'm now at 30M rows - no issues so far.