1、遇到如下提示，这就是MySQL的文件描述不够用了。
SQLSTATE[HY000]: General error: 23 Out of resources when opening file '/tmp/#sql_5c10_0.MYD' (Errcode: 24)
2、Linux(Ubuntu)下MySQL的安装与配置
首先检查系统中是否已经安装了MySQL：sudo netstat -tap | grep mysql
安装Mysql：sudo apt-get install mysql-server mysql-client
测试安装是否成功：sudo netstat -tap | grep mysql
启动MySQL服务： sudo start mysql
停止MySQL服务： sudo stop mysql
修改 MySQL 的管理员密码： sudo mysqladmin -u root password newpassword
设置远程访问(正常情况下，mysql占用的3306端口只是在IP 127.0.0.1上监听，拒绝了其他IP的访问（通过netstat可以查看到）。取消本地监
听需要修改 my.cnf 文件：)：
sudo vi /etc/mysql/my.cnf
bind-address = 127.0.0.1 //找到此内容并且注释
MySQL安装后的目录结构分析(此结构只针对于使用apt-get install 在线安装情况)：
数据库存放目录： /var/lib/mysql/
相关配置文件存放目录： /usr/share/mysql
相关命令存放目录： /usr/bin(mysqladmin mysqldump等命令)
启动脚步存放目录： /etc/rc.d/init.d/
3、Mysql文件描述符命令
查看当前打开了多少个文件描述符：lsof -u mysql|wc -l
因为数据库链接不了，所以只能通过系统查看他的文件描述符（31311可以通过sudo netstat -tap | grep mysql查看）
root@localhost:~# cat /proc/31311/limits
4、修改mysql打开文件限制，参考http://forums.cpanel.net/f354/upgrade-mysql-5-5-sqlstate-hy000-general-error-23-out-resources-286172.html
This likely means that you’ve reached the open files limit for MySQL. Check to see how many open files:
Code:
sudo -u mysql bash ulimit -a | grep open
You’ll likely see something such as this:
Code:
bash-4.1$ ulimit -a | grep open open files (-n) 2048
If that’s the case, exit that user’s login:
Code:
exit
Now, add the following to /etc/security/limits.conf:
Code:
echo -e "mysql soft nofile 10000nmysql hard nofile 20000" >> /etc/security/limits.conf
Check it added the lines to the file:
Code:
cat /etc/security/limits.conf
At that point, log back in as the user and check open files limit again:
Code:
sudo -u mysql bash ulimit -a | grep open
Exit again and now also increase the limit in /etc/mysql/my.cnf（linux版本是my.cnf，一般会放在/etc/my.cnf，/etc/mysql/my.cnf）by editing the file to include this line under the [mysqld] section:
Code:
open_files_limit=20000
Restart MySQL and check it has updated:
Code:
/etc/init.d/mysql restart mysqladmin var | grep open_files_limit
If that doesn’t work, it might be disk space itself, although that error is typically a different one when /tmp has run out of space. You should check the size of /tmp when the error is occurring:
Code:
df -kh | grep /tmp