LMS 4.0 Discovery

[Marvin Rhoads]

OK, I'll take the plunge and post an LMS 4.0 question.

I built a new installation on my Windows 2008 x64 SP1 server (Xeon 5140 CPU, 12 GB RAM, 16 GB swap plenty of disk. Doesn't belong to a domain and AV not running) and chose to go the route of a complete fresh install. I backed up my old data (not a lot but I may choose to import it later), uninstalled LMS 3.2 completely (including removing the CSCOpx directory), and updated my server's browser to Firefox 3.6.10 so I can test locally.

Installation seemed to go fine. No errors or exceptions were reported. I've set up TACACS+ authentication an a 3rd party SSL certificate (more on those later, one question per thread ) but for now am authenticated locally as the admin superuser.

I went through the getting started workflow and set up a default credential set with SNMP credentials confirmed (pasted right out of the live router on my local network). I used that same router (actually a Switch Virtual Interface in a Catalyst 4506) as a seed device, alson with two others for grins to run an initial discovery. I told CW to start discovery and sat back.

It reports "Devices discovered 3, Devices unreachable 3". I've turned on cdp and arp/routing table discovery modules. I have snmp v2c to snmp v1 fallback selected. (My devices are all snmp v1.) All three devices are reachable (via ping and ssh and telnet) from the CW server. There are no ACLs - indeed one of them is my default gateway!

Basically , simple discovery bombed.

Any suggestions for troubleshooting?

[Joe Clarke · Accepted solution]

The default credential sets don't come into play yet.  For discovery to set devices as reachable, you need to make sure you've entered the read-only community string on the SNMP Settings page of the Discovery settings.  See the attached screenshot.

If the devices are still unreachable, a sniffer would come in very handy.  Filter on all SNMP traffic to the seed devices, and see what LMS is using to poll them and how they are responding.

[Marvin Rhoads]

That did it, Joe.Thank you.

I made my target *.*.*.* for my default SNMP RO password and re-ran discovery successfully. I noted the new set of unreachables and iterated against those (some with differing snmp passwords). Adding that second set as specific hosts with their associated passwords cleared those up.

I'm moving out from there, but that's looking much more promising. I'll post separately on a remote access issue.