Hey Jonathan,
Yes, it's possible to run a PRTG Cluster without any Remote Probe(s). However, it is not possible to run sensors on the Failover ONLY while it is possible to run them on the Master only or on both cluster nodes simultaneously.
Please let me explain you how clustering with PRTG works.
After setting up a Failover Node, a "Cluster Probe" appears on the Master Node, so from now on, the Master Node has a Local Probe and a Cluster Probe. At the same time the Failover Node only has a Cluster Probe, but no Local Probe.
The important bit is the Cluster Probe. It actually is a combination of the Local Probes of all the PRTG Cluster Nodes. All devices and sensors that you either add as new objects under the Cluster Probe, or that you move (not clone, not copy) to the "Cluster Probe", will be monitored by the Master Node and by the Failover Node as well. Please simply check this out.
We've designed the architecture like this, that only the Master Node has its Local Probe still shown (but not the Failover Nodes), because there may be some devices, which do not need to be monitored by all the Cluster Nodes, because these devices are not as important.
So let me summaries this again:
- If you want to monitor a device only on the Master Node, please leave it underneath your Local Probe.
- If you want to monitor a device from Master Node and Failover Node as well, please move it underneath the Cluster Probe.
The Cluster Probe represents ALL Cluster Nodes, the Master Node and ALL Failover Nodes in one entity.
So, in case that the Master Node goes down, the Failover Node keeps monitoring all devices underneath the Cluster Probe, but will not monitor the devices underneath your Local Probe.
Of course you can also deploy a device under the Cluster Probe and under the Local Probe as well and deploy different sensors to this (now two) devices.
For details, please also read the following articles:
Best regards,
Sven Roggenhofer [Paessler Technical Support]
Add comment