NPM not seeing backup route - showing down nodes

Each of our locations have a primary T1 and a backup T1. If the primary T1 goes down, NPM will show that down plus all the nodes in that office. The issue is that our network is setup so that everything gets re-routed over the backup T1s automatically. Nodes are still up and pingable and a trace route confirms the path over our backup T1, but NPM still shows nodes as down.

Is this normal or possibly a bug? Running the latest version of NPM.

Thanks

Find more posts tagged with

Accepted answers

BakerD

Ok, I got it figured out. I believe it was a few different issues. One was with only one particular site. If I disabled CEF, it would work. Rebooting the router, clearing/resetting the CEF DB, etc had no effect. I upgraded the IOS and it started working. So either that IOS was buggy or CEF was corrupt somehow. I have other routers running the same IOS still that work fine. I will probably go ahead and put the newer IOS on them anyway.

Another issue I didn't like is that if the T1 link went down, NPM would show the whole router/node as down even though it was just the serial interface. The node was still up and reachable via our backup route. So I changed NPM to monitor the IP on the ethernet interface.

Now I just need to set all my other sites like that.

All comments

Don331

I bet the IP address in the node details is the interface address that's down.

Set a loopback address on the routers and poll from that address.

Don't forget to add these commands to the cisco router config for proper SNMP/Trap reporting:

snmp-server trap-source Loopback0
snmp-server source-interface informs Loopback0

Don

BakerD

I am talking about all my servers, Wireless APs etc in that office that are all still reachable. Even the primary router is reachable via the backup route to its ethernet interface. So really all that should show as down is the office router's serial interface. Since the router can still be reach via the backup route on its ethernet interface I don't think NPM should show the router as down, just that interface. If it couldn't reach any interface on that router then it should show the router as down. That is another issue too. Right now my main concern is all my servers, switches APs. They are not down and can be pinged still but NPM shows them as down.

I bet the IP address in the node details is the interface address that's down.
Set a loopback address on the routers and poll from that address.
Don't forget to add these commands to the cisco router config for proper SNMP/Trap reporting:
snmp-server trap-source Loopback0
snmp-server source-interface informs Loopback0

Don

BakerD

Still having this issue, anyone have any ideas/thoughts?

Thanks

brian.austin

When you confirmed this via ping and traceroute, were you on the NPM poller? And is the NPM poller on the same subnet? I'm curious if the route BACK to the NPM is seen from the other side. That will cause a down issue as well.