MySQL master slave due to ibdata1 full disk space--Master and slave failure
Because innodb_file_per_table is set = 1,ibdata1 is still bursting to fill the disk space
When the master/slave is disconnected, the IO thread is connected and the SQL thread is broken.
Want to know why ibdata1 is growing so much?
The individual understands this:
Master and slave disconnected, IO thread in, get the event things update, and the SQL thread is broken, resulting in a large number of undo, burst Ibdata1.
The final verification found that undo was full of ibdata1.
Download a gadget: py_innodb_page_info.py himself online download link: http://pan.baidu.com/s/1i3fanhr
To create a directory:
[Email protected]]# mkdir InnoDB
Upload the downloaded file to the InnoDB directory
[email protected] innodb]# python py_innodb_page_info.py/export/servers/mysql/data/ibdata1
Total number of page:10395648:
Insert Buffer Free list:3878
Insert Buffer bitmap:632
System page:130
Transaction system Page:1
Freshly allocated page:54247
Undo Log page:10209130
File Segment inode:118869
B-tree node:8129
File Space header:1
Extension description page: 631
Total data page: 10395648 undo Data page: 10209130 accounted for almost 95%.
MySQL ibdata1 Full disk space