Mod-Gearman to bridge high-latency connections?

[ottow]

Hi,

Can Mod-Gearman be used to make checks over high-latency connections more reliable?
We are considering various options for consolidating our NagiosXI infrastructure, and would prefer to have only one single NagiosXI configuration to maintain, rather than as now with several regional systems in various parts of the world.
It is my understanding that both Fusion and Federated requires maintaining regional XI servers and their configuration, whereas if we would use Mod-Gearman we would only have to maintain one single configuration (and regional Mod-Gearman slaves).

Latency is however a concern, on some connections we could see up to 300ms between the NagiosXI server and the Mod-Gearman slaves.
How well will Mod-Gearman work under these conditions?

Thanks,

Otto

[mbellerue]

That is a hard one to tell. I, personally, haven't seen anyone using Mod Gearman with that kind of latency. Will this new model have VPNs between the site with XI and the site with the servers? That may open up the option for passive checks to NRDP.

To keep going down the Mod Gearman route, your license does allow you to run a test instance of XI. You could spin this up and test it out over time without impacting production. It would be hard to test it at load like that, but it would at least get you started collecting data points.

[ottow]

VPN tunnels between sites and the central XI server yes.
300ms is really the very worst-case scenario, it is the highest delay we currently see with our existing infrastructure but in the new design we hope to get it down quite a bit. I doubt we will get below 200ms though.
A large portion of our monitoring is thru SNMP (network infrastructure-type equipment), it is my understanding NDRP can't do that?

It does seem like we need to test this to really find what will work best. Thanks for mentioning the test instance option.

[mbellerue]

Correct, NRDP can't accept SNMP traps. XI can accept them, Admin -> SNMP Trap Interface. But SNMP travels over UDP, and I would worry that with a high latency, you might also have a (relatively) high chance for packet loss. As the old saying goes, "I could tell you a UDP joke, but you might not get it." One of the reasons why I was thinking NRDP, with agents (preferably NCPA) installed on your remote devices is that TCP HTTP are pretty tolerant of iffy connections. But you may find that SNMP works fine as well.

[ottow]

Hi,

Trying to implement mod-gearman now, but am running into an issue when installing the Worker on a CentOS Linux release 8.1.1911 (Core) as per below.
The requested package https://repo.nagios.com/nagios/8/nagios ... noarch.rpm is not available, all I see is an earlier version (nagios-repo-8-1.el8.noarch.rpm)

Verifying nagios repo is installed...CentOS-8 - AppStream 1.9 MB/s | 6.4 MB 00:03
CentOS-8 - Base 1.3 MB/s | 5.0 MB 00:03
CentOS-8 - Extras 624 B/s | 2.1 kB 00:03
not installed.
ERROR: Please install the nagios repo with this command: rpm -Uvh https://repo.nagios.com/nagios/8/nagios ... noarch.rpm
[root@got-modgearman1 tmp]# rpm -Uvh https://repo.nagios.com/nagios/8/nagios ... noarch.rpm
Retrieving https://repo.nagios.com/nagios/8/nagios ... noarch.rpm
curl: (22) The requested URL returned error: 404 Not Found
error: skipping https://repo.nagios.com/nagios/8/nagios ... noarch.rpm - transfer failed
[root@got-modgearman1 tmp]# rpm -Uvh https://repo.nagios.com/nagios/8/nagios ... noarch.rpm
Retrieving https://repo.nagios.com/nagios/8/nagios ... noarch.rpm
curl: (22) The requested URL returned error: 404 Not Found
error: skipping https://repo.nagios.com/nagios/8/nagios ... noarch.rpm - transfer failed
[root@got-modgearman1 tmp]#

[benjaminsmith]

Hello @ottow,

We haven't rolled out support for CentOS 8 quite yet. Can you set this up on a 7 system?

Integrating Mod-Gearman with Nagios XI

[ottow]

Aha, ok that explains it.
Will shift to 7.7

Thanks,

Otto

[optionstechnology]

Is it possible to tell Mod_Gearman which checks to run on which worker?

If its multi-region it needs to run the correct regions checks on each regional worker in this scenario

[benjaminsmith]

Hello,

Yes, this can be controlled in the Mod German configuration file ( module.conf). Take a look at the following KB article as it explains how the queues and workers function.

Nagios XI - Mod-Gearman Queues and Workers

A number of scenarios are described, but creating host or service groups for each region would be one way to control which checks will not be executed on a specific worker.

[ottow]

Hi,

I've switched to CentOS Linux release 7.7.1908 (Core), and am now getting a similar error:

Verifying nagios repo is installed...not installed.
ERROR: Please install the nagios repo with this command: rpm -Uvh https://repo.nagios.com/nagios/7/nagios ... noarch.rpm
[root@got-modgearman1 tmp]# rpm -Uvh https://repo.nagios.com/nagios/7/nagios ... noarch.rpm
Retrieving https://repo.nagios.com/nagios/7/nagios ... noarch.rpm
curl: (22) The requested URL returned error: 404 Not Found
error: skipping https://repo.nagios.com/nagios/7/nagios ... noarch.rpm - transfer failed
[root@got-modgearman1 tmp]# cat /etc/redhat-release
CentOS Linux release 7.7.1908 (Core)
[root@got-modgearman1 tmp]# ^C

Please advise,

Otto

Hello @ottow,

We haven't rolled out support for CentOS 8 quite yet. Can you set this up on a 7 system?

Integrating Mod-Gearman with Nagios XI