How to shutdown a host from Nagios Xi

[Johnjones]

I have a UPS it sends SNMP TRAPS to nagios . If the UPS battery status is critical.I want to initiate a graceful shutdown of the host connected to the UPS from Nagios XI. Can anyone please help me with this?

[hsmith]

Event Handlers are going to be your best friend in making this work. Read up, and let us know what you need help with.

[rkennedy]

You will need a remote agent installed such as NRPE, or NSClient++ installed on the machine you're looking to shut down. It will then need a command defined, that references a bash or windows script to shut down the system.

Then, you'll want to use event handlers which will trigger every time a state changes (OK->CRITICAL,CRITICAL->WARNING). Have this event handler script detect if it's critical, and if it is, then send the command over NRPE to shut it down.

This post has quite a bit of useful information - https://support.nagios.com/forum/viewto ... 999#bottom

[Johnjones]

You will need a remote agent installed such as NRPE, or NSClient++ installed on the machine you're looking to shut down. It will then need a command defined, that references a bash or windows script to shut down the system.

Then, you'll want to use event handlers which will trigger every time a state changes (OK->CRITICAL,CRITICAL->WARNING). Have this event handler script detect if it's critical, and if it is, then send the command over NRPE to shut it down.

This post has quite a bit of useful information - https://support.nagios.com/forum/viewto ... 999#bottom

Thank you very much it really helped. Can you tell if what I am doing is right. I added a shutdown script to the client server and then added a shutdown command to nrpe.cfg referring to that script. This is the shutdown script host_shutdown.

Code: Select all

#host_shutdown
!#/bin/bash
logger  "shutting down the server due to Low UPS battery at `date`" 
sudo shutdown -h now 

Then I added this event handling script to the nagios server.

Code: Select all

#!/bin/bash

SERVICESTATE=$1 # UPS state
HOSTADDRESS=$2 # host address which is connected to the UPS

if [ "$SERVICESTATE" == "CRITICAL" ] ; then
      /usr/local/nagios/libexec/check_nrpe -H $HOSTADDRESS -c host_shutdown
    fi
exit 0

[rkennedy]

It looks right, from the Nagios machine can you run /usr/local/nagios/libexec/check_nrpe -H $HOSTADDRESS -c host_shutdown (replace $HOSTADDRESS$ with the IP of the machine that you setup the machine that you're looking to restart.)

Does that work for you?

One thing I didn't see mentioned, did you define the command for the event handler, and assign it to the UPS service?

[Johnjones]

It looks right, from the Nagios machine can you run /usr/local/nagios/libexec/check_nrpe -H $HOSTADDRESS -c host_shutdown (replace $HOSTADDRESS$ with the IP of the machine that you setup the machine that you're looking to restart.)

Does that work for you?

i tried running it manually but the output is this

Code: Select all

[root@localhost ~]# /usr/local/nagios/libexec/check_nrpe -H xxx.xx.x.xxx -c host_shutdown
NRPE: Unable to read output

[rkennedy]

This may not be relevant as there isn't any echo in your bash, nor a 'exit 0' -- did the server actually restart? What happens if you modify the shutdown bash script to this?

Code: Select all

#host_shutdown
!#/bin/bash
logger  "shutting down the server due to Low UPS battery at `date`"
echo "shutting down"
sudo shutdown -h now
exit 0 

[Johnjones]

I tried the script you posted but, same output. It didn't work and the server didn't shutdown.

I'm running NRPE using xinetd so I added the host_shutdown command to the nrpe.cfg. how to restart nrpe to take this in to effect?

[rkennedy]

Ah - try 'service xinetd restart' (since it's running under xinetd). Take a look at your /var/log/messages for an error message on the client side when you attempt to restart from the server using the check_nrpe command.

[Johnjones]

I gave nagios sudo access to the client server and changed the file permissions for host_shutdown script to nagios:nagios in client server and it worked. I was able to shutdown the server.