Hello.
I'm confused about the Distributed Monitoring Architecture and can't find much detailed documentation on it.
I plan on having 1 master node and up to 6 child nodes. The master node will be at operations center and will only monitor the child nodes and their hosts.
For my simple test setup, I'm using a master and a single child node. I've done this setup two different ways.
The first way, I set up hosts using a combination of the master server frontend as well as the child node frontend. Operationally speaking, it worked as expected. However, when creating maps, images seemed to be duplicated (as if the child node sent images to the master node -- I can only assume that if I had 6 child nodes that I would have 7 copies of each image (6 from the child nodes and 1 from the master). I'm pretty sure that other things (value maps?) were copied in this way.
The second attempt, I recreated the master (same node id though). I assumed that the child node would send everything (templates, hosts, custom value maps, etc) to the master. They are apparently talking however I can't see any child node hosts. I can see in the logs where the servers are sending configurations and history though.
The only other way to configure it I can think of is to use the recreate the child node and configure it entirely from the master node frontend. Is this how it should be done?
Ultimately, all child nodes will be remote unmanned sites. The master node server will be shipped a the same time the first child node system (including server) will. I'd like to configure the master node server and child node servers at my office to provide a turnkey solution. However I won't have all of the child node servers by the time the master node server is shipped. Is this possible? And if, so, how do I prevent the duplication of common things like the images?
Thanks for your time and consideration. Any help is appreciated?
I'm confused about the Distributed Monitoring Architecture and can't find much detailed documentation on it.
I plan on having 1 master node and up to 6 child nodes. The master node will be at operations center and will only monitor the child nodes and their hosts.
For my simple test setup, I'm using a master and a single child node. I've done this setup two different ways.
The first way, I set up hosts using a combination of the master server frontend as well as the child node frontend. Operationally speaking, it worked as expected. However, when creating maps, images seemed to be duplicated (as if the child node sent images to the master node -- I can only assume that if I had 6 child nodes that I would have 7 copies of each image (6 from the child nodes and 1 from the master). I'm pretty sure that other things (value maps?) were copied in this way.
The second attempt, I recreated the master (same node id though). I assumed that the child node would send everything (templates, hosts, custom value maps, etc) to the master. They are apparently talking however I can't see any child node hosts. I can see in the logs where the servers are sending configurations and history though.
The only other way to configure it I can think of is to use the recreate the child node and configure it entirely from the master node frontend. Is this how it should be done?
Ultimately, all child nodes will be remote unmanned sites. The master node server will be shipped a the same time the first child node system (including server) will. I'd like to configure the master node server and child node servers at my office to provide a turnkey solution. However I won't have all of the child node servers by the time the master node server is shipped. Is this possible? And if, so, how do I prevent the duplication of common things like the images?
Thanks for your time and consideration. Any help is appreciated?