mysql无法正常封闭问题清查mingxing - 娱乐之横扫全球

mysql无法正常封闭问题清查mingxing

2019-02-11 11:51:58 | 作者: 傲菡 | 标签: 清查,问题,函数 | 浏览: 8942

bin/mysqld_safe: line 137: 15569 Segmentation fault      nohup /data/mysql5/libexec/mysqld –defaults-file=my.cnf –based1100919 15:53:40 mysqld_safe Number of processes running now: 0100919 15:53:40 mysqld_safe mysqld restarted

 

阐明:正常的履行mysqladmin应该输出如下

100919 15:50:18 mysqld_safe mysqld from pid file /data/mysql5/run/mysqld.pid ended

 

2、清查

       刚开始以为是代码问题,改用debug形式时发现在履行进程3)的时分mysqld就dump导致重启了。清查导致dump的代码履行途径如下

add_index – row_merge_build_indexes – row_merge_sort –  row_merge –  row_merge_blocks_copy

    前两个函数在handler/handler0alter.cc,后边的函数在row/row0merge.c,功用是新建的索引排序并合并到同一个文件中。

       发现dump的点居然是row_merge_blocks_copy函数的传参进程,因而怀疑是内存拜访方面的问题。运用valgrind试运行。

在非debug形式下运用valgrind发动mysql,发现在履行shutdown指令时,valgrind有如下输出:

5998 Thread 16:5998 Invalid read of size 45998    at 0×30B560B0F0: pthread_kill (in /lib64/libpthread-2.5.so)5998    by 0×7822A4: thr_alarm (in /home/dingqi/innodb_ssd/mysql/libexec/mysqld)

 

5998    by 0×598E62: my_real_read(st_net*, unsigned long*) (in /home/dingqi/innodb_ssd/mysql/libexec/mysqld)

5998    by 0×599264: my_net_read (in /home/dingqi/innodb_ssd/mysql/libexec/mysqld)

5998    by 0×5AA14E: handle_one_connection (in /home/dingqi/innodb_ssd/mysql/libexec/mysqld)

5998    by 0×30B56062F6: start_thread (in /lib64/libpthread-2.5.so)

5998    by 0×30B4AD1B6C: clone (in /lib64/libc-2.5.so)

5998  Address 0×3197D9D0 is on thread 15’s stack

版权声明
本文来源于网络,版权归原作者所有,其内容与观点不代表娱乐之横扫全球立场。转载文章仅为传播更有价值的信息,如采编人员采编有误或者版权原因,请与我们联系,我们核实后立即修改或删除。

猜您喜欢的文章