Too many open files in system

One: Problem description

Received the message to the police, said a server MySQL is not connected.

I tried it manually, connected for more than half an hour, MySQL has not been connected, and no feedback error messages.

Error log errors:

161128 9:50:18 [ERROR] Error in Accept:too many the open files in system

161128 9:54:34 [ERROR] Error in Accept:too many the open files in system

When you switch to a MySQL user, you also report an error:

[Root@vhost_eip_8 init.d]# Su-mysql

SU:/bin/bash:too many open files Insystem

I tried to restart MySQL, and the result was a reboot failed.

[Root@vhost_eip_8 mysql]# Service Mysqldrestart

Shutting down MySQL ..... Determine

Starting mysql.the Server quit withoutupdating PID file (/[failed]ysql/vhost_eip_8.pid).

Error log errors:

161128 12:29:09 Mysqld_safe mysqld from Pidfile/data/mysql/vhost_eip_8.pid ended

161128 12:29:10 Mysqld_safe starting Mysqlddaemon with databases From/data/mysql

^g/usr/local/mysql/bin/mysqld:file './mysql-bin.index ' not Found (errcode:23)

161128 12:29:10 [ERROR] Aborting

161128 12:29:10 [Note]/usr/local/mysql/bin/mysqld:shutdown complete

The error of the newspaper is really Bishi, Mysql-bin.index document exists clearly.

I found many processes on the monitor server that connect to the server, such as:

Root 28897 28655 0 12:08? 00:00:00/usr/local/mysql/bin/mysql--host=10.0.3.47--port=3306--user=monitor--password=x xxxxxxxxx-e showdatabases;

Root 29479 29245 0 10:44? 00:00:00/usr/local/mysql/bin/mysql--host=10.0.3.47--port=3306--user=monitor--password=x xxxxxxxxx-e showdatabases;

Root 29890 29653 0 11:31? 00:00:00/usr/local/mysql/bin/mysql--host=10.0.3.47--port=3306--user=monitor--password=x xxxxxxxxx-e showdatabases;

Root 29973 29731 0 11:59? 00:00:00/usr/local/mysql/bin/mysql--host=10.0.3.47--port=3306--user=monitor--password=x xxxxxxxxx-e showdatabases;

Root 30139 29932 0 11:03? 00:00:00/usr/local/mysql/bin/mysql--host=10.0.3.47--port=3306--user=monitor--password=x xxxxxxxxx-e showdatabases;

And I have every minute of the monitoring of the scheduled task to cancel, the result is still receiving an email to the police.

Oh, my heart.

I killed all these processes on the monitor server:

Ps-ef | grep MySQL | grep-v grep |cut-c 9-15 |xargs kill-9

Then suddenly received more than 10 alarm mail, I was drunk, this is a flash flood outbreak? ：

After a few seconds, no more to receive the alarm mail. Two: The cause of the error

The Fs.file-max value setting is too small.

This value can be viewed in this way:

Cat/proc/sys/fs/file-max

Found the result was 65536.

III: Solutions

Modify the value Fs.file-max in the/etc/sysctl.conf to 6553600

Then execute the SYSCTL-P

At this point, and then restart MySQL, it started successfully.

--The main reference of this article: Https://zhidao.baidu.com/question/179045384319146284.html,

Http://blog.sina.com.cn/s/blog_8da9c942010116vn.html