Critical Path Configuration

derek_lineman · September 13, 2019, 4:43pm

Hey everyone,

Even though Critical Path configuration I’ve setup seems to show correctly (info > path outages), I’m not getting the “pathOutage” event only the individual nodeDown events.

I’ve setup the Foreign Parent ID on each node in order for to be shown on the path outages screen. The Parent ID is in a separate provisioning request so I provided the name to that as well.

Also, I’ve set pathOutageEnabled=“true” in poller-configuration.xml and issued a pollerd restart request then restarted openNMS completely when it didn’t seem to do anything.

From what I’ve read in the docs, the nodeDown events should still be generated but they should have additional “eventReason” parameters along with the seperate “pathOutage” event.

Update: nodeDown event description does state ‘All interfaces on node x.x.x.x are down because of the following condition: pathOutage’ but still don’t see the pathOutage event.

Prior to use the provisioning method, I’d also setup the Critical Paths manually through Notifications > Configure Path Outages and the Primary Interface IP but this didn’t seem to work either.

Currently running Horizon 24.1.3.

In case it matters, I’ve been testing the outage by blackholing the routes on the openNMS server itself rather than “killing” my real network, for example:

ip route add blackhole x.x.x.x

Would anyone be able fill me in on any other configuration I may have missed?

Thanks

Derek

derek_lineman · September 28, 2019, 10:30am

No one got any ideas?

Apologies if this is a simple one but I’m completely stumped.

Jonathan · February 5, 2021, 2:19pm

Hi @derek_lineman

Did you find your problem? I’m having the same issue with vers 27.0.2. I have a Node for my internet connection, a node for my VPN and all other node are under VPN. Path Outage display correctly, but when the LINK node is going down, I’m receiving alert for all node except only for the top one.

tomminerd · March 10, 2021, 10:54pm

I’m seeing the same issue in 27.1.0. It has to be something simple. I’m going to try restarting my OpenNMS server completely and see what that does.

UPDATE

I got it working after I set the Parent Foreign source on one of my nodes since its parent node was in a different requisition.

I have TOR switches acting as the parent node for all nodes in the rack. Those are in the same requisition so no need to provide a parent foreign source. The TOR switch is connected to a distribution switch which is in a different requisition so I just specified that in the TOR node Parent Foreign Source. Once I did that, I tested email notifications and only got a notification for the TOR switch.