CLOSE_WAIT和TIME_WAIT处理

早上登陆服务器的时候输入netstat -an|grep mysql

发现存在大量TIME_WAIT状态的连接

tcp        0      0 127.0.0.1:3306              127.0.0.1:41378             TIME_WAIT
tcp        0      0 127.0.0.1:3306              127.0.0.1:41379             TIME_WAIT

tcp        0      0 127.0.0.1:3306              127.0.0.1:39352             TIME_WAIT

tcp        0      0 127.0.0.1:3306              127.0.0.1:39350             TIME_WAIT
tcp        0      0 127.0.0.1:3306              127.0.0.1:35763             TIME_WAIT
tcp        0      0 127.0.0.1:3306              127.0.0.1:39372             TIME_WAIT
tcp        0      0 127.0.0.1:3306              127.0.0.1:39373             TIME_WAIT
tcp        0      0 127.0.0.1:3306              127.0.0.1:41176             TIME_WAIT

通过调整内核参数解决

vi /etc/sysctl.conf

编辑文件，加入以下内容：
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_tw_reuse = 1
net.ipv4.tcp_tw_recycle = 1
net.ipv4.tcp_fin_timeout = 30

然后执行/sbin/sysctl -p让参数生效。

net.ipv4.tcp_syncookies = 1表示开启SYN Cookies。当出现SYN等待队列溢出时，启用cookies来处理，可防范少量SYN攻击，默认为0，表示关闭；

net.ipv4.tcp_tw_reuse = 1表示开启重用。允许将TIME-WAIT sockets重新用于新的TCP连接，默认为0，表示关闭；

net.ipv4.tcp_tw_recycle = 1表示开启TCP连接中TIME-WAIT sockets的快速回收，默认为0，表示关闭。

net.ipv4.tcp_fin_timeout修改系統默认的TIMEOUT时间

修改之后，再用命令查看TIME_WAIT连接数

netstat -ae|grep “TIME_WAIT” |wc –l

   发现大量的TIME_WAIT 已不存在，mysql进程的占用率很快就降下来的，网站访问正常。

不过很多时候，出现大量的TIME_WAIT状态的连接，往往是因为网站程序代码中没有使用mysql.colse()，才导致大量的mysql  TIME_WAIT.

 

{
sysctl -w net.ipv4.tcp_keepalive_time=30
sysctl -w net.ipv4.tcp_keepalive_probes=2
sysctl -w net.ipv4.tcp_keepalive_intvl=2
sysctl -w net.ipv4.tcp_fin_timeout=30
sysctl -w net.ipv4.tcp_syncookies=1
sysctl -w net.ipv4.tcp_tw_reuse=1
sysctl -w net.ipv4.tcp_tw_recycle=1
}

 

 

What It Means
"Proto" is short for protocol, which is either TCP or UDP. "Recv-Q" and "Send-Q" mean receiving queue and sending queue. These should always be zero; if they're not you might have a problem. Packets should not be piling up in either queue, except briefly, as this example shows:
tcp 0 593 192.168.1.5:34321 venus.euao.com:smtp ESTABLISHED
That happened when I hit the "check mail" button in KMail; a brief queuing of outgoing packets is normal behavior. If the receiving queue is consistently jamming up, you might be experiencing a denial-of-service attack. If the sending queue does not clear quickly, you might have an application that is sending them out too fast, or the receiver cannot accept them quickly enough.
"Local address" is either your IP and port number, or IP and the name of a service. "Foreign address" is the hostname and service you are connected to. The asterisk is a placeholder for IP addresses, which of course cannot be known until a remote host connects. "State" is the current status of the connection. Any TCP state can be displayed here, but these three are the ones you want to see: