Nagios Support Forum

I apologize if this is a question from left field. We are getting ready to deploy a 2nd node to our Nagios Log Server (soon to be a cluster). I was told (the person shall remain nameless to protect the innocent) that there is a log server manager
that would sit above the cluster similar to Fusion. I have searched the documentation and find no reference to it so now I find myself having to ask the question.

Thanks,

You can manage the cluster from any of the NLS instances. Any of them can technically be the 'manager'. Can I get a little more information on what you mean by manage the cluster? Or correct me if I am understanding this question wrong.

What I had heard was it would be like Fusion. From Fusion you can see and manage the different Nagios Servers. From this management server you would manage the different nodes in the cluster.
I don't have a lot of information, only asking the question because of what I had heard. The visio above shows how it was explained.
Again I know I may be out in left field because I have found no documentation but wanted to make sure.

If you connect more than one instance to your Nagios Log Server installation, the Web GUI of _any_ node can be used to view the entirety of your cluster, including all log data, instance status, configuration, etc. You can access any instance and have a consistent result - many client sit a load balancer or similar in front of all of the cluster members for this reason.

Nice to see you again @gregwhite!

Thank you. I think this is cleared up for me. From any node in the cluster, using the web gui we can see the status of the cluster. The load balancing is something outside of NLS. If we want to use a load
balancer, we should use something like a netscaler.

Appreciate your help and patience.
Greg

I'm pretty unfamiliar with third party load balancers. To the best of my understanding, NLS does a pretty good job balancing its own load when you have multiple instances, I'm not sure what the interaction would be like if you were to add another one on top of that.

Ok. J. Olson mentioned that some clients put a load balancer in front of the log server instances. However if NLS does a good job on its own load balancing it seems best not to complicate the configuration.
Also at this point we only have two instances.

Thanks for you help.
Greg

I'll get clarification from Jesse about that when I can. We sit in the same room he's just out this morning.

Greg,

While Nagios Log Server does load balance the _data_ that you input, it cannot load balance the incoming information.

What this means is that there is a possibility that clients could send all of their log data to a single node in the cluster. After the data is done arriving, it will be distributed in the backend. The data can _only arrive at that single node_. This means that if that single node goes down, all of your logs will stop entering the cluster.

When people set up round-robin DNS/a netscaler or similar, they do so between their log senders and their Nagios Log Server instances, so that if one node goes down, all of their logs continue coming in.

The second thing a load balancer allows you to do is use HTTP/HTTPS load balancing between your nodes, since they all share the same GUI. This allows you to reduce the stress of a particular servers apache daemon.

Does the above make sense? Please let me know if you need any clarification.

Jesse

Thank you for all your help!! I think I have what I need now. I will follow up if I need any further clarification.

Have a great weekend.