Jeff fair question and the answer is yes
1. We have many customers who use polling engines to monitor duplicate IP space. So we require a method to differentiate nodes and we do this by poller
2. With 10.0 you can use the ignore feature to ignore those same nodes on discovery on one of the polling engines
Brandon,
We're just beginning to experiment with Network Sonar Discovery in v10. We have the same situation as Jeff. We have 3 pollers. For the sake of this discussion, we have a Class B network 10.1.0.0 / 16 with all the nodes split across all 3 pollers (based on a custom property). We would like to be able to set up a daily scheduled discovery to discover any new nodes. Am I hearing right, we would need to create and schedule separate discoveries for each poller? How would we use the ignore feature based on response #2 if the ignore list is global? From what I can determine the ignore list is global and the nodes are not individually tied/paired to the specific discovery job.
Hi,
Some quick answers. Let me know if I can elaborate.
1. You don't need to schedule discoveries for each poller if you can ping the devices from a single poller. However, if you will want to import any of the new nodes, then you will want them discovered on the corresponding poller. It sounds like you might want to depending on how you've split up your nodes across your 3 pollers.
2. The ignore list is per engine ID (poller) not by discovery. Hence, you can ignore different IP addresses, interfaces, and volumes for different pollers.
Thanks
Do you think it is practical for everyone with additional pollers to use scheduled discovery? The answer is simple. The discovery does not work its supposed to be! You should improve it.
You have that kind of bugs because you still did not implement multiple IP support for single node which requested by a lot of customers for a long time.
Also I have the similar issue.
Last week we added our additional poller. But, like others our discoveries are now useless to me. When it discovers everything twice it makes the list unmanageable. I don't feel that the proper solution is for me to have to ignore everything that has been discovered again simply because it is on the other poller.
Can someone fix the software to make a discovery option available so it can determine if the device is being monitored by any of the engines and if not being monitored then show it as a new device? I can see how some customers may like having each poller do its own discovery and show all new devices that each poller finds. That is easy to do if a poller is responsible for an entire subnet. But not all customers work that way.
Is this still on-going or has SolarWinds fixed the issue?
Installed our new poller last Friday and setup a network sonar scan and its finding duplicates.
Also is there some type of automatic load balancing for Pollers?
Thanks
The duplicate nodes issue between pollers has not changed in 10.2 with regards to the discussion in this thread.
There is not automatic load balancing for pollers currently in the Orion Platform products.
Thanks
How are people handling when a device has been discovered by a poller and another poller had already discovered the device? Are they just ignoring?
