mod gearman

[elinagios]

Hello

i dont fully understand mod gearman after going through the documentation, so few questions about that.

1) Once i install mod gearman on a functional Nagios XI server, does it mean all the checks will be go through the mod gearman "queue"?
2) If i disable the mod gearman worker on the Nagios XI server while there are no other mod gearman workers, will all the checking stop?
3) Is it possible so that Nagios core does most of the checks and for some sites(subnets) i use mod gearman workers? For this as i understand need to use hostgroups and servicegroups so that mod gearman worker understands which checks it needs to run.
4) The worker, you can install it on linux and nothing else needed, beside the plugins? No nagios server or such?

Thank you
Mirgo

[cdienger]

1) Yes. By default all checks will be sent to the "host" or "service" queues.

2) Yes. Without workers in place nothing will be able to run by default.

3) To have core run most of the checks, the checks should belong to the host and/or service group specified by the localhostgroups= and localservicegroups= options in modules.config

4) Correct. Just the worker and Nagios plugins are needed.

[elinagios]

3) Is it possible to avoid mod gearman process once it is installed on NagiosXI server? So that core can run checks via the old method and checks coming from some mod gearman worker, will go through mod gearman?
5) Once mod gearman is installed, then all the new checks should be added to service/hostgroups defined in the mod gearman conf, else the checks will not work, correct?
6) If one would decide he doesnt want to use mod gearman anymore but just to use core, then just stop service, core takes over and that's all?

[scottwilkerson]

3) Is it possible to avoid mod gearman process once it is installed on NagiosXI server? So that core can run checks via the old method and checks coming from some mod gearman worker, will go through mod gearman?

This will completely stop checks form going through mod_gearman
Just comment this line out of the nagios.cfg

Code: Select all

broker_module=/usr/lib/mod_gearman2/mod_gearman2.o config=/etc/mod_gearman2/module.conf eventhandler=no

to

Code: Select all

#broker_module=/usr/lib/mod_gearman2/mod_gearman2.o config=/etc/mod_gearman2/module.conf eventhandler=no

And restart nagios

Code: Select all

service nagios restart

5) Once mod gearman is installed, then all the new checks should be added to service/hostgroups defined in the mod gearman conf, else the checks will not work, correct?

This is only if you want to segregate which worker runs which checks, otherwise all workers can run all checks. What is described is an advanced setup.

6) If one would decide he doesnt want to use mod gearman anymore but just to use core, then just stop service, core takes over and that's all?

See my answer to #3 above

[elinagios]

7) What is Mod Gearman capability running host and service checks. I currently have Nagios XI server, core running all the checks. Installing Mod Gearman on this nagiosxi server will make this server also the worker. I have 3900 host and 3000 service checks running in 5minute intervals. Nagios Core handles them without problem, now when all checks will go through Mod Gearman, will it still be able to handle the load?

[scottwilkerson]

7) What is Mod Gearman capability running host and service checks. I currently have Nagios XI server, core running all the checks. Installing Mod Gearman on this nagiosxi server will make this server also the worker. I have 3900 host and 3000 service checks running in 5minute intervals. Nagios Core handles them without problem, now when all checks will go through Mod Gearman, will it still be able to handle the load?

Yes it would handle the load (potentially much more than your XI server alone), but unless you have a reason, I would be cautious about adding mod_gearman just because. By doing so you are adding additional items/server to fail. If your server is handling the load fine, unless you have another really compelling reason to use mod_gearman I wouldn't until you need to.

Most of our clients don't look into using such a solution until the need to perform over 20,000 checks on a 5 minute interval.

[elinagios]

Have one more question, hopefully last for this topic:
8) After installing mod gearman service on Nagios XI server, it also becomes the worker. What if the Worker proccess dies? The host/service updates wount end up in the monitoring view and i guess no emails are sent out as well. In this case, should i install another instance of NagioXI that monitors the production environment?

[scottwilkerson]

Have one more question, hopefully last for this topic:
8) After installing mod gearman service on Nagios XI server, it also becomes the worker. What if the Worker proccess dies? The host/service updates wount end up in the monitoring view and i guess no emails are sent out as well. In this case, should i install another instance of NagioXI that monitors the production environment?

You can and should add multiple workers if you are going to add mod_gearman, this way you can have mod_gearman workers die and everything keeps on going.

It is also a good idea to monitor your monitoring server, and we do recommend this as well, so much so that we have a Wizard created to monitor other Nagios XI servers.

[elinagios]

So from last post i got 2 questions:
1) I plan to add multiple workers and yes, i can monitor the workers from the worker that is on nagiosxi server. But what if the worker on nagiosxi dies or worse, the mod-gearman server service. I should monitor worker on nagiosxi server with other workers? What about the mod-gearman server service? If that dies, i guess all the workers cant forward info anyways.
2) The secondary nagiosxi - this doesnt need to be licensed if i keep the hosts below 7 ?

Mirgo

[scottwilkerson]

1) The local worker is no different than the remote worker. If it dies the others keep working. However, you can monitor all of this from the second Nagios XI

2) correct.