Some pointers may be invalid and cause the dump to abort.
Query (xxxxxxxxx): alter table tb1 add primary key(`col1`,`col2`,`col3`)
Connection ID (thread ID): 129443
Status: NOT_KILLED
......
2017-05-25T14:10:24.858719+08:00 0 [Note] InnoDB: Ignoring data file './dbname/tb0.ibd' with space ID 375, since the redo log references ./dbname/tb0.ibd with space ID 280.
2017-05-25T14:10:24.860628+08:00 0 [Note] InnoDB: Ignoring data file './dbname/tb1.ibd' with space ID 374, since the redo log references ./dbname/tb1.ibd with space ID 279.
......
2017-05-25T14:10:44.635655+08:00 0 [Note] InnoDB: Ignoring data file './dbname/#sql-7b4e1_9f3a1.ibd' with space ID 375. Another data file called ./dbname/tb0.ibd exists with the same space ID.
......
2017-05-25T14:10:39.845947+08:00 0 [Note] InnoDB: Ignoring data file './dbname/#sql-7b4e1_9f3a1.ibd' with space ID 374. Another data file called ./report_gamein/t30741_hs_g47_day.ibd exists with the same space ID.
2017-05-25T14:11:40.547210+08:00 0 [Note] InnoDB: Rollback of trx with id 3377056858 completed
2017-05-25T14:11:40.554385+08:00 0 [Note] InnoDB: Rollback of non-prepared transactions completed
......
2017-05-25T14:11:03.752024+08:00 0 [Warning] Recovery from master pos 43851595 and file binlog.006933 for channel 'amaster'. Previous relay log pos and relay log file had been set to 43851818, /home/mysql/log/relaylog/relaylog-amaster_3537.020527 respectively.
......
2017-05-25T14:11:03.817050+08:00 0 [Warning] Recovery from master pos 789680988 and file binlog.027468 for channel 'bmaster'. Previous relay log pos and relay log file had been set to 789681243, /home/mysql/log/relaylog/relaylog-bmaster_3637.020768 respectively.
......
2017-05-25T14:11:04.353863+08:00 0 [Note] /usr/sbin/mysqld: ready for connections.
Version: '5.7.17-log' socket: '/home/mysql/data/mysqld.sock' port: 3306 MySQL Community Server (GPL)
2017-05-25 14:11:04 0x7f10762de700 InnoDB: Assertion failure in thread 139708678924032 in file fsp0fsp.cc line 2108
InnoDB: Failing assertion: frag_n_used > 0
......
截取出来的是第一次遇到问题, mysql重启进行Crash Recovery的日志;
可以从红色标记的地方看到, 其实第一次重启的时候, crash recovery就已经完成了, 而且mysqld进程也已经ready for connections,
但是马上就触发了同样的问题, 导致mysql又发生了Crash, 而且mysqld_safe也跟着"消失"了;
在之后的重启尝试中, 每次在Crash之前, 都有这么一行信息:
点击(此处)折叠或打开
2017-05-25T15:25:33.025244+08:00 0 [Note] InnoDB: Cleaning up trx with id 3377057419
结合堆栈信息中显示出来的fsp0fsp.cc line 2108, 在源代码中找到这一行(红色标注),