DNS Server Monitor Bug

Hello.

I'm hoping someone can help me as I'm struggling.

I have BIND9 running on two machines that can talk to each other. One is the master the other, the slave. They're working perfect.

However, every 5 minutes in my logs named is started.

I have narrowed it down to the Monitor script which is executing

PHP:

$fp = fsockopen('udp://localhost',53, $errno, $errstr, 2);

So I made this script, ran it and I get...

PHP:

<?php
$fp = fsockopen('udp://localhost',53, $errno, $errstr, 2);
echo $errno;
echo '---#' . $errstr;
echo '--#' . $fp;
?>

Code:

0---#--#Resource id #4

Monitor script cannot connect to localhost port 53 however, I can Telnet to it correctly.

I assume it's something to do with PHP but I have no idea what! All new named services are by the user 'nobody'.

As you can imagine running multiple named processes means the DNS hangs on both services which makes redundancy a major issue. Help!

 

Hi Till. It is, but every 5 minutes it's restarting.

I've turned OFF ISPConfig monitoring but it's still restarting.

Here is the Array of rescue_module_serverconfig.ser.txt

Code:

Array
(
[serverconfig] => Array
(
[server] => Array
(
[loglevel] => 2
)

[rescue] => Array
(
[try_rescue] => y
[do_not_try_rescue_httpd] => y
[do_not_try_rescue_mysql] => y
[do_not_try_rescue_mail] => n
)

)

)

It gets overwritten everytime the cron runs.

 

Hi Till.

Even when I turn off monitoring and in the database try_rescue=n and the file 'rescue_module_serverconfig.ser.txt' has the correct values it's still restarting BIND.

 

Any ideas as to why this is happening?

At the moment every 3 hours I've got a cron set to 'killall named' then Monit restores the service.

 

Hi Till.

Turns out it wasn't you at all - it was Munin-Node starting BIND every 5 minutes. Still no idea why!

Thanks for the great product - I'm still in awe of it! Going on our 28th server tomorrow.