[Mysql] Mysql slave restarts at regular intervals

Background: the slave replication process stops every once in a while from the HuaYun_Interface-SLAVE slave library, and zabbix sends alarm messages.

Mysql > show slave status\ G

* * 1. Row *

Slave_IO_State:

Master_Host: *

Master_User: repl

Master_Port: 3307

Connect_Retry: 60

Master_Log_File: mysql-bin.000195

Read_Master_Log_Pos: 629275501

Relay_Log_File: mysql-relay-bin.000521

Relay_Log_Pos: 625430128

Relay_Master_Log_File: mysql-bin.000195

Slave_IO_Running: No

Slave_SQL_Running: No

Check the mysql error log when stopping, and we find a clue.

2016-06-23 14:06:56 32289 [Warning] 'proxies_priv' entry' @ root@sh275' ignored in-skip-name-resolve mode.

160623 15:03:49 mysqld_safe Number of processes running now: 0

160623 15:03:49 mysqld_safe mysqld restarted-found that the mysql service was restarted, and we configured skip_slave_start (prevents the repl process from automatically restarting), so the replication process is in the down state

2016-06-23 15:03:50 0 [Warning] TIMESTAMP with implicit DEFAULT value is deprecated. Please use-explicit_defaults_for_timestamp server option (see documentation for more details).

2016-06-23 15:03:50 0 [Note] / home/data/mysql/bin/mysqld (mysqld 5.6.24-log) starting as process 22569.

2016-06-23 15:03:50 22569 [Note] Plugin 'FEDERATED' is disabled.

2016-06-23 15:03:50 22569 [Note] InnoDB: Using atomics to ref count buffer pool pages

2016-06-23 15:03:50 22569 [Note] InnoDB: The InnoDB memory heap is disabled

2016-06-23 15:03:50 22569 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins

2016-06-23 15:03:50 22569 [Note] InnoDB: Memory barrier is not used

2016-06-23 15:03:50 22569 [Note] InnoDB: Compressed tables use zlib 1.2.3

2016-06-23 15:03:50 22569 [Note] InnoDB: Using Linux native AIO

2016-06-23 15:03:50 22569 [Note] InnoDB: Not using CPU crc32 instructions

2016-06-23 15:03:50 22569 [Note] InnoDB: Initializing buffer pool, size = 10.0G

2016-06-23 15:03:51 22569 [Note] InnoDB: Completed initialization of buffer pool

At the same time, we check the email sent by zabbix, and there is a linux host memory alarm notification.

Alarm host: _ HuaYun_Interface-SLAVE

Host IP: *

Alarm time: 15:08:06 on June 23, 2016

Alarm level: Average

Alarm message: Lack of available memory on server HuaYun_Interface-SLAVE

Question details: Available memory:163.15 MB

Event ID: 42590

Log on to the mainframe for inspection

[root@NEW-APP-MYSQL-SLAVE ~] # dmesg

Out of memory: Kill process 32289 (mysqld) score 966 or sacrifice child

Killed process 32289, UID 498, (mysqld) total-vm:14630648kB, anon-rss:7711284kB, file-rss:360kB [root@NEW-APP-MYSQL-SLAVE] # free-m

Total used free shared buffers cached

Mem: 7872 7747 124 0 124 1331

-/ + buffers/cache: 6292 1580

Swap: 2047 56 1991

It is obvious that there is an OOM killer phenomenon caused by insufficient memory in linux. There is no way to increase the memory of the host.