The curse of Dennis D. Spreen

Hi there, I’ve been trying to set this up to monitor Lighttpd but I’ve been experiencing this weird issue where if I run the command locally as one user, I get the correct output. Any subsequent runs by other users get no output. What’s worse, trying to get the results through NRPE yields a “NRPE: Unable to read output” from check_nrpe.

Do you run this check locally or remotely?

Hi There,
I’ve tried to use the check_apachestatus_auto.pl Plugin, but it always fails.
The Check is:
./check_apachestatus_auto.pl –hostname “hostname” –url http://“hostname”/server-status?auto
If I tried it with user root and user nagios out of the bash, all works fine.

If i tried to check it via nrpe or online via nagios, I always becomes a failure:

(Service check did not exit properly)

My output in the bash is:
qsargos01:/opt/nagios/libexec # ./check_apachestatus_auto.pl –hostname qsargos01.total.hosting –url http://argos-test.total.hosting/server-status?auto
APACHE OK – 0.028 sec. response time, Busy/Idle 1/74, open 949/1024, ReqPerSec 0.1, BytesPerReq 1289, BytesPerSec 111|Idle=74 Busy=1 OpenSlots=949 Slots=1024 Starting=0 Reading=0 Sending=1 Keepalive=0 DNS=0 Closing=0 Logging=0 Finishing=0 ReqPerSec=0.086761 BytesPerReq=1289 BytesPerSec=111.918873 Accesses=154qsargos01:/opt/nagios/libexec #

It looks like there is a Problem with the output data. Why there is always this line “qsargos01:/opt/nagios/libexec #” directly behind the output?

Thanks in advance

Hello,

Thank you for this update. May I suggest a new change ?

*** mine.pl 2010-08-05 15:34:54.374819435 +0200
— check_apachestatus_auto.pl 2009-07-14 13:04:30.000000000 +0200
*************** use Digest::MD5 qw(md5 md5_hex);
*** 30,36 ****
# Nagios specific

use lib “/usr/local/nagios/libexec”;
- use lib “/usr/lib/nagios/plugins”;
use utils qw(%ERRORS $TIMEOUT);
#my %ERRORS=(‘OK’=>0,’WARNING’=>1,’CRITICAL’=>2,’UNKNOWN’=>3,’DEPENDENT’=>4);

— 30,35 —-

This patch allow your script to work on a debian installation, using debian packages.

Hello again,

For the lib, it’s not a problem – perl -I is a good friend.

But I checked how you get Server type… and, well, I think it can be done in a smarter way:

diff -u check_apachestatus_auto.pl check_webserver.pl
— check_apachestatus_auto.pl 2009-07-14 13:04:30.000000000 +0200
+++ check_webserver.pl 2010-08-06 08:58:17.000000000 +0200
@@ -190,6 +190,9 @@
my $webcontent = undef;
if ($response->is_success) {
$webcontent=$response->content;
+ my @t = split(‘/’, $response->header(‘server’)) ;
+ $httpserver = $t[0];
+

my $Uptime = 0;
if($webcontent =~ m/Uptime: (.*?)\n/) {
@@ -212,11 +215,8 @@
}

my $BusyWorkers= 0;
- if($webcontent =~ m/(BusyWorkers|BusyServers): (.*?)\n/) {
- $BusyWorkers = $2;
- if ($1 eq ‘BusyServers’) {
- $httpserver = ‘LIGHTTPD’;
- }
+ if($webcontent =~ m/(:?BusyWorkers|BusyServers): (.*?)\n/) {
+ $BusyWorkers = $1;
}

my $IdleWorkers=0;

I’m curious if you have any interest in expanding the functionalities of this plugin. We’re looking to monitor the activity of threads on a more detailed level. We have some situations where after an incident we’re stuck with tons of threads in the ‘Sending Reply’ or ‘Graceful Finishing’ state. I’m curious if you would consider adding in functionality to watch thread types and allow for warning and critical thresholds to be applied to them.

Thanks!!

sorry, overlooked your comment – those thread infos are hard to retrieve (you need to write an extra apache module).. I am not aware of any existing modules to use – do you know any?

I am getting a error with the PNP file.

templates/check_apachestatus.php [30]:

Undefined index: Slots

And I really have no clue how to fix this.

never mind — helps if I name the file right.

You’re welcome.

Sorry just saw that you replied. There are no extras needed to provide the thread scoreboard. It’s part of server status. All this plugin would need to do is parse the scoreboard and generate notifications based on the number of threads in a certain state.

Thx for your really helpfull apache check.

we have added 2 additional features:
– once added the path to make your check work on debian
– changed the warn/critical from absolut values to percent (as many apache are configured in a way to have a different slot count after the first apache master process generation)

Both formatted as nice and seperated patches:
https://github.com/gebi/nagios_checks/commit/c7519cc7e00ba54a489d043984ef3ae102527287

https://github.com/gebi/nagios_checks/commit/a2fd9f4f05bca68881a8e67ba1c939c025865098

what do you think about them?

A note to people who get the “NRPE: Unable to read output” error. The script leaves a tempfile in /tmp that is owned by the user who runs it. So if you run it as a user other than that which runs your NRPE it will die because it can’t create a new dropfile.

To fix this, delete the file (it’s called /tmp/_check_apachestatus_auto. and then add the line “unlink($TempFile); to line 289 of the script, right before the “http://www.tcomm.es/joomla/index.php/en/downloads-menu/finish/19-unix/11-checkunixmemusage/0.html” line.

Thx Charles, fixed my problem.
Propesed change to the script

# my $TempFile = $TempPath.$o_host.’_check_apachestatus_auto’.md5_hex($url);
#The $ is the effective user id.
my $name=getpwuid($<);
my $TempFile = $TempPath.$o_host.'_check_apachestatus_auto'.'_'.$name.'_'.md5_hex($url);