[Nagiosplug-help] mysqllag plugin

[Jonathan Angliss]

On 05/01/11 20:20, Tim Dunphy wrote:
> Hello list!! I am attempting to implement a plugin that alerts when it detects lag in mysql repliction. It seems really useful and I would like to get this going, but I am running into an issue in doing so.
>
> I am receiving this error in the nagios ineterface:
>
>
> ## nagios script error
>
> MySQL Lag
> 	
> Notifications for this service have been disabled
> 	CRITICAL 	05-01-2011 16:19:58 	0d 4h 55m 58s 	3/3 	(Return code of 127 is out of bounds - plugin may be missing)

====8<-------
snipping host info
====8<-------

> ## locally mysqllag check script works! :)

What does "locally" mean? On the Mysql server? Or is this check_mysqllag 
command on the nagios box?

> [root at virtcent10:/usr/local/nagios/libexec] #./check_mysqllag
> + STATE_OK=0

=====8<-----
snip plugin output
=====8<-----

> ## ls -l of script on nagios client
>
>
> [root at virtcent10:/usr/local/nagios/libexec] #ls -l check_mysqllag*
> lrwxrwxrwx 1 nagios nagios   17 May  1 11:12 check_mysqllag ->  check_mysqllag.sh
> -rwx------ 1 nagios nagios 1908 May  1 13:20 check_mysqllag.sh

virtcent10 is the client machine, while virtcent11 is the nagios server 
right?

> ## this is the mysqllag service definition
>
> define service{
>          use                             generic-service         ; Name of service template to use
>          host_name                       db2
>          service_description             MySQL Lag
>          check_command                   check_mysqllag
>          notifications_enabled           0
>          }
>
>
> ## this is the mysqllag command definition
>
> define command{
>          command_name    check_mysqllag
>          command_line    $USER1$/check_mysqllag -H $HOSTADDRESS$ -p 3306 -v $ARG1$ $ARG2$
>          }
>

How does Nagios execute this command on the remote host? From the output 
above, it looks like you put the plugin on the remote host, but Nagios 
is trying to run it locally. This kind of execution requires the use of 
an intermediary agent, such as nrpe, or the check_by_ssh plugin. 
Otherwise, Nagios is actually trying to execute the script in the LOCAL 
(to nagios) path, which according to your output, may not be where your 
script has it.

> ## this is the mysqllag script
>
>
> #! /bin/sh -x
>
> STATE_OK=0
> STATE_WARNING=1
> STATE_CRITICAL=2
> STATE_UNKNOWN=3
> STATE_DEPENDENT=4
> mysqlpath='/usr/bin/'
> user='root'
> pass='secret'
> warn=30
> crit=60
> null="NULL"
> usage1="Usage: $0 -u$user -p$pass [-w<warn>] [-c<crit>]"
> usage2="<warn>  is lag time, in seconds, to warn at.  Default is 30."
> usage3="<crit>  is lag time, in seconds, to be critical at.  Default is 60."
>
> exitstatus=$STATE_WARNING #default
> while test -n "$1"; do
>      case "$1" in
>          -c)
>              crit=$2
>              shift
>              ;;
>          -w)
>              warn=$2
>              shift
>              ;;
>          -u)
>              user=$2
>              shift
>              ;;
>          -p)
>              pass=$2
>              shift
>              ;;
>          -h)
>              echo $usage1;
> 	    echo
>              echo $usage2;
>              echo $usage3;
>              exit $STATE_UNKNOWN
> 	    ;;
> 	-H)
>              host=$2
>              shift
>              ;;
>          *)
>              echo "Unknown argument: $1"
>              echo $usage1;
> 	    echo
>              echo $usage2;
>              echo $usage3;
>              exit $STATE_UNKNOWN
>              ;;
>      esac
>      shift
> done

Your command_line definition is passing in a -v, but you have no handle 
for it, which means it'll fall to the *), and as such may always return 
an error that you have an unknown argument.  Any reason that's there?

> thank you in advance for any suggestions you might have to get this working!!

My recommendation would be to copy the check_mysqllag script onto the 
Nagios host.  Change your command definition to look like this:

define command {
     command_name   check_mysqllag
     command_line   $USER1$/check_mysqllag -H $HOSTADDRESS$ -u $ARG1$ \
			-p $ARG2$ -w $ARG3$ -c $ARG4$
}

I've wrapped the command_line for easy readability, it should be on one 
line.

Your service definition would then look like this:

define service {
     .... usual other svc stuff .....
     check_command	check_mysqllag!username!password!30!60
}

The alternative to above is leaving the plugin where it is, and using 
NRPE, and defining a remote command.  Nagios would use the following:


define command {
     command_name    check_mysqllag
     command_line    $USER1$/check_nrpe -H $HOSTADDRESS$ \
			-c check_mysqllag
}

Then on the mysql side, you'd have an nrpe configuration:

command[check_mysqllag]=/usr/local/nagios/libexec/check_mysqllag -w 30 -c 60

Hope that gives you some ideas.

-- 
Jon Angliss