New user trying to have a worker dedicated to LAN network.

[hallux4]

Hi,

This time, I don't have any problem except my lack of understanding.

I have an OMD setup with a public IP. And I would like to monitor some of our intranet (no port forwarding possible, and dynamic IP).

So I setup a mod-gearman worker on one of our LAN server.

• add the secret key.
• modified the job server IP in worker.cfg.
• set (Distributed Monitoring): -- hostgroups=my_intranet_hostgroup -- eventhandler=no -- notifications=no -- services=no -- hosts=no

On the job server I can see the worker. If I stop the worker on OMD. The jobs are just "Waiting".

If I set services and hosts = yes on the "LAN" worker. it will handle all the jobs, not just the one from hostgroups=my_intranet_hostgroup.

How to make the worker take only its hostgroups jobs ?

I saw this in the documentation, can it be a solution for my problem ?

• queue_custom_variable https://github.com/sni/mod_gearman#how-to-set-queue-by-custom-variable

I am sorry to bother you so often ^^

[hallux4]

I tried like this but I don't think it's working ^^

On OMD server:

define host {
  host_name                      Intranet-Health
  alias                          Intranet Health
  address                        MY_IP
  use                            generic-host
  _WORKER                        hostgroup_test
}

On OMD server:

Queue Name                | Worker Available | Jobs Waiting | Jobs Running
----------------------------------------------------------------------------
 check_results             |               2  |           0  |           0
 eventhandler              |              11  |           0  |           0
 grafana_data              |               0  |          38  |           0
 host                      |               6  |           0  |           0
 hostgroup_hostgroup_test  |               6  |           0  |           0
 hostgroup_linux-servers   |               5  |           0  |           0
 hostgroup_test            |               0  |           2  |           0
 hostgroup_websites_health |               5  |           0  |           0
 notification              |               5  |           0  |           0
 perfdata                  |               1  |           0  |           0
 service                   |               6  |           0  |           0
 worker_node1              |               1  |           0  |           0
 worker_node2              |               1  |           0  |           0
 worker_node3              |               1  |           0  |           0
----------------------------------------------------------------------------

On LAN server (node3):

hostgroups=hostgroup_test

[sni]

it should work exactly how you did it. Both ways are possible. If you stop the LAN worker, you should see job pile up in that hostgroup queue. Using custom variables is possible to give you complete flexibility but its not required in your scenario.

The key options are described here: https://labs.consol.de/nagios/mod-gearman/sample_distributed.png

As mentioned in the "Distributed Monitoring" scenario on https://labs.consol.de/nagios/mod-gearman/index.html#_common_scenarios

[hallux4]

Oh I did in fact missed a step from the picture. I forgot to define the hostgroups in etc/mod-gearman/server.cfg.

Thanks again. Have a great day !