managed interfaces and sonar network discovery

Hi,

I'm new to NPM so be kind to my tender soul when slamming me for putting up a silly post.

Q1. In NPMv10.1 is there a way to permanently put an interface/s in an unmanaged state. Currently when trying to unmanage an interface/s it asks for a "Unmanage this device until" details (screen capture 1). I want to be able to select a bunch of interfaces and click unmanage and forget about it until there is a need to start managing/collecting stats for those interfaces. The way it currently behaves is the unmanaged nodes/interfaces will automatically start being managed again once the timer elapses.

Q2. In sonar network discovery for the entire 10.0.0.0/8 keeps failing/timing out. I have tried the discovery with both default settings (screen capture 2) and with maximising the timers (screen capture 3) without success, am I doing something wrong?

Q3. To get NPM to discovery devices I have had to break the class A subnet into smaller class C chunks, while this works it is not really practical for a network of our size (700+ devices). Question is, for devices with multiple IP addresses is there a way I can force network discovery to discover devices using the management (loopback) address? Or do I have to manually change the source addresses afterwards?

Q4.Following on from Q3. Presuming a router with ip address 1.1.1.1 was discovered and I changed the source ip address afterwards to 2.2.2.2 and ran the discovery again. Will the routers source address revert back to 1.1.1.1?

Q5. Finally how to a configure device discovery to run say in 2hours time or every sunday at 2am? Currently options are to run every x hours or daily or immediately.

Find more posts tagged with

many_interfaces

manage_nodes

management

interface_polling

npm_10

discovery_hangs

interface

manage

sonar_network_discovery

manage_devices

Accepted answers

All comments

karim1

With attachment!

SCREEN CAPTURE.zip

needle

Hi,

Q1 - Try this - go to Admin, Manage Nodes, maximise the node in question, then tick all of the interfaces that you don't want to monitor. Then click on 'Edit Interface' and tick the box that says 'Display interface as unplugged rather than down'.

Q3 - If a discovery is finding systems with multiple IP addresses, the simplest thing to do would be post discovery mark the alternative address and click Ignore. Then correct the other interface if necessary and it shouldn't bother you again.

Q4 - I'm not 100% sure but as above, if you mark it as Ignore in the discovery results it shouldn't change again.

Q5 - Not sure if it's possible - the only thing I can suggest is setting a Custom job to run every x hours, say 168 hours for once a week. Only caveat is kicking the original run off at 2am on a Sunday... not ideal! Someone else might be able to suggest something.

Regards

Dave

karim1

Apologies for the late reply (I was seconded elsewere and just getting back to Orion).

I tired your solution for Q1 however when I look at 'Interface Elements' under Admin->Polling Engine the number counter remains the same. Shouldn't this number decrease once a selection of interfaces have been unmanaged?

karim1

One more thing. Following on from Q1, once an interface has been selected as 'unplugged rather than down', if I drill down on the device and re-select the same interface the 'unmanage' button is highlighted, shouldn't the 'unmanage' button remain greyed out because the interface has already been selected as 'unplugged rather than down'?

This is all very confusing and none of it is documented.

Like I said in my original post what I want to do is select interface/s -> and permanently tell NPM to NOT manage/poll/collect stats for those interfaces. And once this is done have this reflected in the number interface/s being managed (which I am guessing is 'Interface Elements'.

Vineet

MarieB

Hi Karim--

Could you tell me what is missing in the documentation? I'll do some research and see if I can hunt it down--otherwise, I'll let the writer know.

Thanks,