Recently, the mysql server often crashes because of what, although it is logical that due to lack of memory, but this has never happened before. It seems that the server is not loaded, everything is optimized as much as possible. Is that a change of OS to CentOS from Debian. Although Debian has never crashed, more precisely once, but after turning off InnoDB everything was fine. Tell me what to increase in the config, what to reduce? What actions to take to mysql in the end did not fall so often. I would be very grateful to someone who tells you what to do at such a moment, how to optimize the work of MySQL. Here are the actual mysql logs:
160814 15:29:02 mysqld_safe mysqld restarted 160814 15:29:02 [Note] /usr/libexec/mysqld (mysqld 5.5.50-MariaDB) starting as process 610 ... 160814 15:29:02 [Note] Plugin 'InnoDB' is disabled. 160814 15:29:02 [ERROR] mysqld: Out of memory (Needed 128917504 bytes) 160814 15:29:02 [ERROR] mysqld: Out of memory (Needed 96681984 bytes) 160814 15:29:02 [ERROR] mysqld: Out of memory (Needed 72499200 bytes) 160814 15:29:02 [Note] Plugin 'FEEDBACK' is disabled. 160814 15:29:02 [Note] Server socket created on IP: '0.0.0.0'. 160814 15:29:02 [Note] Event Scheduler: Loaded 0 events 160814 15:29:02 [Note] /usr/libexec/mysqld: ready for connections. Version: '5.5.50-MariaDB' socket: '/var/lib/mysql/mysql.sock' port: 3306 MariaDB Server 160814 15:29:04 [ERROR] mysqld: Table './database/table' is marked as crashed and should be repaired 160814 15:29:04 [Warning] Checking table: './database/table'
Configuration file my.cnf:
[mysqld] default-storage-engine=myisam skip-innodb local-infile=0 innodb_file_per_table = 1 pid-file = /var/run/mysqld/mysqld.pid datadir=/var/lib/mysql socket=/var/lib/mysql/mysql.sock # Disabling symbolic-links is recommended to prevent assorted security risks symbolic-links=0 query-cache-size = 8192 thread-cache-size = 4 [mysqld_safe] log-error=/var/log/mariadb/mariadb.log pid-file=/var/run/mariadb/mariadb.pid