I run a rather busy site, and during peek hours I see over 10.000 open connections to my database server on my webserver when a run a netstat command. 99% of the connections are in the TIME_WAIT state.

I learned about this mysql variable: wait_timeout http://dev.mysql.com/doc/refman/5.1/en/server-system-variables.html#sysvar_wait_timeout today. Mine is still set at the default 28.800 seconds.

Is lowering this value safe?

Non of my queries usually takes over a second. So it seems silly to keep a connection open for 480 minutes.

I also heard about using mysql_pconnect instead of mysql_connect, but i've been reading nothing but horror stories about it, so I think i'll stay away from that.

  • 3
    There's a difference between queries and connections. You need to at least ensure your web site software won't break if a shorter wait_timeout cause a connection to close when the software expects it to remain open. Commented Feb 1, 2012 at 22:15

3 Answers 3


Lowering the value is pretty trivial without a mysql restart

Let's say you want to lower timeouts to 30 seconds

First, add this to my.cnf


Then, you can do something like this

mysql -uroot -ppassword -e"SET GLOBAL wait_timeout=30; SET GLOBAL interactive_timeout=30"

All DB Connections after this will timeout in 30 seconds


Make sure to use explicitly use mysql_close. I do not trust Apache as most developers do. If not, sometimes, there is a race condition where Apache closes a DB Connection but does not inform mysqld and mysqld hold that connection open until it times out. Even worse, you may see TIME_WAITs more often. Choose your timeout values wisely.

UPDATE 2012-11-12 10:10 EDT


After applying my posted suggestions, create a script called /root/show_mysql_netstat.sh with the following lines:

netstat | grep mysql > /root/mysql_netstat.txt
cat /root/mysql_netstat.txt | awk '{print $5}' | sed 's/:/ /g' | awk '{print $2}' | sort -u > /root/mysql_netstat_iplist.txt
for IP in `cat /root/mysql_netstat_iplist.txt`
        ESCOUNT=`cat /root/mysql_netstat.txt | grep ESTABLISHED | awk '{print $5}' | grep -c "${IP}"`
        TWCOUNT=`cat /root/mysql_netstat.txt | grep TIME_WAIT   | awk '{print $5}' | grep -c "${IP}"`
        IPPAD=`echo "${IP}..................................." | cut -b -35`
        (( ESCOUNT += 1000000 ))
        (( TWCOUNT += 1000000 ))
        ES=`echo ${ESCOUNT} | cut -b 3-`
        TW=`echo ${TWCOUNT} | cut -b 3-`
        echo ${IPPAD} : ESTABLISHED:${ES} TIME_WAIT:${TW}
echo ; echo
netstat -nat | awk '{print $6}' | sort | uniq -c | sort -n | sed 's/d)/d/'

When you run this, you should see something like this:

[root@*** ~]# /root/ShowConnProfiles.sh : ESTABLISHED:00002 TIME_WAIT:00008 : ESTABLISHED:00000 TIME_WAIT:00002 : ESTABLISHED:00001 TIME_WAIT:00000 : ESTABLISHED:00000 TIME_WAIT:00079 : ESTABLISHED:00002 TIME_WAIT:00001 : ESTABLISHED:00003 TIME_WAIT:01160 : ESTABLISHED:00002 TIME_WAIT:00000 : ESTABLISHED:00000 TIME_WAIT:00589 : ESTABLISHED:00001 TIME_WAIT:00570

      1 established
      1 Foreign
     11 LISTEN
   1301 TIME_WAIT

If you still see a lot of mysql TIME_WAITs for any given web server, here are two escalation steps to take:


Login to the offending web server and restart apache as follows:

service httpd stop
sleep 30
service httpd start

If necessary, do this to all the web servers

service httpd stop (on all web servers)
service mysql stop
sleep 120
service mysql start
service httpd start (on all web servers)


You can force the OS to kill TIME_WAITs for mysql or any other app with the following:

echo ${SEC_TO_TIMEWAIT} > /proc/sys/net/ipv4/tcp_tw_recycle
echo ${SEC_TO_TIMEWAIT} > /proc/sys/net/ipv4/tcp_tw_reuse

This will make TIME_WAITs time out in 1 second.

To give credit where credit is due...

  • 1
    Nearly 2k views for this and only 2 up votes (including mine, so 1), wtf?! This is a great answer. Sorry I can only give 1 up vote!
    – Baldrick
    Commented Dec 20, 2012 at 15:04
  • I tried this technique, and I had a lot of CLOSE_WAIT connections (2622) but not TIME_WAIT connections. How should I interpret this result? Commented Feb 12, 2013 at 11:28
  • Escalation #2 is dangerous!! serverfault.com/a/715883
    – Freedo
    Commented Jan 27, 2021 at 7:04

If you are getting a lot of TIME_WAIT connections on the MySQL Server then that means that MySQL server is closing the connection. The most likely case in this instance would be that a host or several hosts got on a block list. You can clear this by running:

mysqladmin flush-hosts

To get a list of the number of connections you have per IP run:

 netstat -nat | awk {'print $5'} | cut -d ":" -f1 | sort | uniq -c | sort -n

You can also confirm this is happening by going to one of your clients that is having trouble connecting and telnet to port 3306. It will show a message with something like:

telnet mysqlserver 3306
Connected to mysqlserver.
Escape character is '^]'.
sHost 'clienthost.local' is blocked because of many connection errors; unblock with 'mysqladmin flush-hosts'Connection closed by foreign host.

If you have lots of TIME_WAIT connections to your MySQL server, it means you code is running lots of queries against your DB, and opening/closing a connection for each query.

In this case, you should use persistent connectivity to your DB server, using the MySQLi extension.


If you can't use MySQLi, you should instead use the thread_cache_size parameter in your MySQL configuration.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .