Service check timed out after 60.04 seconds

[dgatlas]

I've written a plugin in php which I can run at the command line fine.

When run against time, here are the values. It takes less than 5 seconds to run:

Code: Select all

real	0m2.144s
user	0m0.099s
sys	0m0.015s

But... in the XI interface I get: (Service check timed out after 60.04 seconds). When I increase the nagios service_check_timeout to 120 i get: (Service check timed out after 120.05 seconds)

[Box293]

I've written a plugin in php which I can run at the command line fine.

Are you running the plugin as the nagios user?

Code: Select all

su nagios
./my_plugin --arguements blah

[dgatlas]

Are you running the plugin as the nagios user?

Code: Select all

su nagios
./my_plugin --arguements blah

Yes. Running the script as the nagios user also produces the desired output:

Code: Select all

# su nagios
$ ./my_plugin -H 192.168.100.100 -S
Running
$

[Box293]

At this point I suggest commenting out chunks of code in your plugin and then forcing a check from Nagios. Kind of like drawing a line in the sand.

This will help diagnose if a part of the plugin is causing the issue.

[dgatlas]

At this point I suggest commenting out chunks of code in your plugin and then forcing a check from Nagios. Kind of like drawing a line in the sand.

This will help diagnose if a part of the plugin is causing the issue.

It's throwing the error at:

Code: Select all

$result = curl_exec($curl);

Here's the entire method. I don't see anything that could be causing this. I have the same line in another method in the same script that does not cause this issue.

Code: Select all

function getSystem($url, $cookiejar) {

    $curl = curl_init();
    curl_setopt_array($curl, array(
        CURLOPT_RETURNTRANSFER  => true,
        CURLOPT_HEADER          => false,
        CURLOPT_FOLLOWLOCATION  => true,
        CURLOPT_SSL_VERIFYPEER  => false,
        CURLOPT_ENCODING        => "",
        CURLOPT_USERAGENT       => 'spider',
        CURLOPT_AUTOREFERER     => true,
        CURLOPT_CONNECTTIMEOUT  => 30,
        CURLOPT_TIMEOUT         => 30,
        CURLOPT_MAXREDIRS       => 10,
        CURLOPT_COOKIEFILE      => $cookiejar,
        CURLOPT_COOKIEJAR       => $cookiejar,
        CURLOPT_URL             => $url

    ));

    $result = curl_exec($curl);
    curl_close($curl);

    return $result;
}

[dgatlas]

Now the plugin seems to run but shows

Code: Select all

(No output returned from plugin)

Also when defining the service, there is a "Test Check Command" button, which also outputs correctly.

Code: Select all

COMMAND: /usr/local/nagios/libexec/my_plugin.php -H 192.168.100.100 -S
OUTPUT: Running

The plugin was uploaded through the XI Admin page > Manage Plugins

[Box293]

I'm not familiar with curl but I've been looking online at some examples.

I know this is probably not going to make a difference, but perhaps define the array first.

Code: Select all

function getSystem($url, $cookiejar) {

	$options_array = array(
		CURLOPT_RETURNTRANSFER  => true,
		CURLOPT_HEADER          => false,
		CURLOPT_FOLLOWLOCATION  => true,
		CURLOPT_SSL_VERIFYPEER  => false,
		CURLOPT_ENCODING        => "",
		CURLOPT_USERAGENT       => 'spider',
		CURLOPT_AUTOREFERER     => true,
		CURLOPT_CONNECTTIMEOUT  => 30,
		CURLOPT_TIMEOUT         => 30,
		CURLOPT_MAXREDIRS       => 10,
		CURLOPT_COOKIEFILE      => $cookiejar,
		CURLOPT_COOKIEJAR       => $cookiejar,
		CURLOPT_URL             => $url
	);

	$curl = curl_init();
	curl_setopt_array($curl, $options_array);

	$result = curl_exec($curl);
	curl_close($curl);

	return $result;
}

Is there a specific error that is being returned?

Also, are CURLOPT_COOKIEFILE and CURLOPT_COOKIEJAR supposed to be using the same variable $cookiejar?

Code: Select all

      CURLOPT_COOKIEFILE      => $cookiejar,
      CURLOPT_COOKIEJAR       => $cookiejar,

[Box293]

Also when defining the service, there is a "Test Check Command" button, which also outputs correctly.

Be wary of using the Test Check Command button as those commands are executed as the user apache instead of nagios and some funny character escaping can occur. Use the "force an immediate check" on the service screen for a true test.

[dgatlas]

I've removed

Code: Select all

CURLOPT_COOKIEJAR       => $cookiejar

already, that was just for debug. The script seems to pass all that after I changed:

Code: Select all

CURLOPT_CONNECTTIMEOUT  => 30,
CURLOPT_TIMEOUT         => 30,

to

Code: Select all

CURLOPT_CONNECTTIMEOUT  => 10,
CURLOPT_TIMEOUT         => 10,

Now it doesn't display any output:

Code: Select all

(No output returned from plugin)

If I change the exit code to

Code: Select all

exit(1);

That change is reflected on the dashboard with a "Warning" status but still no output.

[dgatlas]

Be wary of using the Test Check Command button as those commands are executed as the user apache instead of nagios and some funny character escaping can occur. Use the "force an immediate check" on the service screen for a true test.

Running as "nagios" still outputs correctly. Both using

Code: Select all

su nagios

and

Code: Select all

sudo -u nagios -H ./my_plugin ...

and yes I'm using "Schedule a forced immediate check".