Comments on: Debugging sleeping connections with MySQL http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/ Everything about MySQL Performance Sat, 21 Nov 2009 05:23:57 -0800 http://wordpress.org/?v=2.8.4 hourly 1 By: Tom Pittlik http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-564403 Tom Pittlik Thu, 21 May 2009 00:12:24 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-564403 I should also add - increasing the max_connections (mysql) and MaxClients (apache) variables to high values is not a solution! In fact it can just prolong the bottleneck as the server takes longer to crash. I should also add – increasing the max_connections (mysql) and MaxClients (apache) variables to high values is not a solution! In fact it can just prolong the bottleneck as the server takes longer to crash.

]]> By: Tom Pittlik http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-564400 Tom Pittlik Thu, 21 May 2009 00:09:50 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-564400 I was seeing this occurring due to every apache connection opening a mysql connection (as is common in a PHP driven website). If there is a bottleneck, a vicious cycle occurs: 1. apache requests increase 2. mysql queries increase 3. mysql slow queries increase 4. apache connections start to wait on slow mysql queries 5. apache connections take longer to close 6. more apache connections are opened 7. with each apache request a new sleeping mysql connection is made 8. mysql & apache reach max_connections 9. server slows to a crawl Solution: - Reduce the time a mysql connection is open during script execution - don't open a mysql connection in the header and close it in the footer. One approach is to create a query wrapper function and open a connection to the database only when the first query occurs - e.g: function query ($query) { if (!mysql_connection) mysql_connect(); return mysql_query(); } - Reduce slow queries (always a good strategy!) I was seeing this occurring due to every apache connection opening a mysql connection (as is common in a PHP driven website). If there is a bottleneck, a vicious cycle occurs:

1. apache requests increase
2. mysql queries increase
3. mysql slow queries increase
4. apache connections start to wait on slow mysql queries
5. apache connections take longer to close
6. more apache connections are opened
7. with each apache request a new sleeping mysql connection is made
8. mysql & apache reach max_connections
9. server slows to a crawl

Solution:
- Reduce the time a mysql connection is open during script execution – don’t open a mysql connection in the header and close it in the footer. One approach is to create a query wrapper function and open a connection to the database only when the first query occurs – e.g:
function query ($query) {
if (!mysql_connection) mysql_connect();
return mysql_query();
}
- Reduce slow queries (always a good strategy!)

]]> By: Mike http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-564338 Mike Wed, 20 May 2009 23:04:13 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-564338 what the hell, here's the first part of the php script that this thing cut off due to bad html stripping it looks like: $link = mysql_connect ("localhost", "root", "password"); $query = "show processlist;"; $result = mysql_query($query) or die ("failed $query"); while ($row = mysql_fetch_array($result)) { if (($row['Time'] > what the hell, here’s the first part of the php script that this thing cut off due to bad html stripping it looks like:

$link = mysql_connect (”localhost”, “root”, “password”);

$query = “show processlist;”;

$result = mysql_query($query) or die (”failed $query”);

while ($row = mysql_fetch_array($result))
{
if (($row['Time'] >

]]> By: Mike http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-564335 Mike Wed, 20 May 2009 23:02:43 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-564335 A better way than restarting your server or whatever is to set a MySQL value that does not allow connections to last over 30 min, or you can have a cron script that checks the db and kills really old sleeping connections. my.cnf: wait_timeout=180 # = 30min php cron script: #!/usr/bin/php -q 1000 && $row['Command'] == "Sleep") && ($row['User'] != "Root")) { $killed = 0; mysql_query("kill ".$row['Id']) or ($killed = 1); if ($killed = 1) { echo "Killed process number ".$row['Id'].", with a User of ".$row['User'] ." in database ".$row['db'].", which had a time of ".$row['Time'] .", and a command of ".$row['Command']."\r\n\r\n"; } } } @mysql_close($link); ?> A better way than restarting your server or whatever is to set a MySQL value that does not allow connections to last over 30 min, or you can have a cron script that checks the db and kills really old sleeping connections.

my.cnf:
wait_timeout=180 # = 30min

php cron script:

#!/usr/bin/php -q
1000 && $row['Command'] == “Sleep”) && ($row['User'] != “Root”))
{
$killed = 0;
mysql_query(”kill “.$row['Id']) or ($killed = 1);
if ($killed = 1)
{
echo
“Killed process number “.$row['Id'].”, with a User of “.$row['User']
.” in database “.$row['db'].”, which had a time of “.$row['Time']
.”, and a command of “.$row['Command'].”\r\n\r\n”;
}
}

}

@mysql_close($link);

?>

]]> By: Richard http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-538914 Richard Tue, 14 Apr 2009 17:48:35 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-538914 Russell, sleeping threads are a problem. They occur when a connection is requested, but is not closed- the connection is still reserved by that thread and "waiting for the client to send a new statement to it". If the sleep time of a thread exceeds the `wait_timeout` variable for mysql, the connection gets closed and recycled. Until then, it is locked. Russell, sleeping threads are a problem. They occur when a connection is requested, but is not closed- the connection is still reserved by that thread and “waiting for the client to send a new statement to it”. If the sleep time of a thread exceeds the `wait_timeout` variable for mysql, the connection gets closed and recycled. Until then, it is locked.

]]> By: Russell G. http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-500621 Russell G. Mon, 09 Mar 2009 20:58:27 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-500621 Are you sure that sleeping threads are a problem? I always thought that a thread showing a Command value of "Sleep" just meant that it had finished whatever request it was working on and was sitting idle, waiting for another request to be assigned to it. I believe the Time value for those threads is how long it's been sitting idle. The MySQL docs seem to say the same thing: "Sleep - The thread is waiting for the client to send a new statement to it." From: http://dev.mysql.com/doc/refman/5.1/en/thread-commands.html Even the program "mytop", which displays the current commands being processed by the server, ignores threads with the Sleep command, I believe for the same reason. I can see how a thread with a Command of "Query" could be a problem if it has a large Time value, but I don't think Sleep commands are a problem. Or have I just completely misunderstood things? Russell G. Are you sure that sleeping threads are a problem? I always thought that a thread showing a Command value of “Sleep” just meant that it had finished whatever request it was working on and was sitting idle, waiting for another request to be assigned to it. I believe the Time value for those threads is how long it’s been sitting idle. The MySQL docs seem to say the same thing:

“Sleep – The thread is waiting for the client to send a new statement to it.”

From:
http://dev.mysql.com/doc/refman/5.1/en/thread-commands.html

Even the program “mytop”, which displays the current commands being processed by the server, ignores threads with the Sleep command, I believe for the same reason.

I can see how a thread with a Command of “Query” could be a problem if it has a large Time value, but I don’t think Sleep commands are a problem. Or have I just completely misunderstood things?

Russell G.

]]> By: Web Services — Troubleshooting Sleeping Processes in MySQL http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-354409 Web Services — Troubleshooting Sleeping Processes in MySQL Thu, 11 Sep 2008 14:18:41 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-354409 [...] Debugging sleeping connections with MySQL [...] [...] Debugging sleeping connections with MySQL [...]

]]> By: Tom Pittlik http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-184670 Tom Pittlik Mon, 05 Nov 2007 00:58:58 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-184670 I get this once in a blue moon - instead of tracking down what's causing the problem (since it's so rare) i made a simple cronjob that checks if dbconnect() fails for mysql_errno() reason 1040 (too many connections). If it does i run shell_exec("httpd restart") and the sleeping connections are gone. I get this once in a blue moon – instead of tracking down what’s causing the problem (since it’s so rare) i made a simple cronjob that checks if dbconnect() fails for mysql_errno() reason 1040 (too many connections). If it does i run shell_exec(”httpd restart”) and the sleeping connections are gone.

]]> By: peter http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-57819 peter Thu, 22 Feb 2007 09:32:43 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-57819 Guruprasad, I'd check if these connections are associated with open file descriptors. If they are it is likely these are persistent connections which may be reused. If not it is likely these connections are still closing by kernel, sometimes it takes longer time for some reason. Guruprasad,

I’d check if these connections are associated with open file descriptors. If they are it is likely these are persistent connections which may be reused. If not it is likely these connections are still closing by kernel, sometimes it takes longer time for some reason.

]]> By: Guruprasad http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/comment-page-1/#comment-57523 Guruprasad Thu, 22 Feb 2007 01:24:58 +0000 http://www.mysqlperformanceblog.com/2007/02/08/debugging-sleeping-connections-with-mysql/#comment-57523 Just in case if some one is tryin this on BSD box, netstat -ntp might not work, u can use lsof instead like, lsof -i -P | grep OR Just in case if some one is tryin this on BSD box,
netstat -ntp might not work, u can use lsof instead like,

lsof -i -P | grep OR

]]>