Solved: LMS management ip & VRF

KAROLY KOHEGYI · ‎10-25-2010

Dear Clarke,

I've read the following URL which was very useful.

https://supportforums.cisco.com/docs/DOC-9005#Preferred_Management_IP

Dsepite of all I have not found solution our problem.

We use the loopback method for discovery which is very good for remote routers. We use many VRF's also in devices. Same devices (etc. switches ) have not loopback interfaces.

So here the LMS choosed the highest IP address of devices . The problem that highest ip address are in the seperated VRF. After discovery the RME can not handle these devices. ( No config archive for example )

Why use the LMS the ip addreses from VRF's?

It will be good for us if LMS uses the "default" VRF ip addresses only. Is it possible???

What is the good, single solution for both routers and switches?

Thanks!

Joe Clarke · ‎10-27-2010

Sorry, I missed your reply. The code that checks for reachability uses SNMP to fetch the sysObjectID. If this fails, the next address in the ipAddrTable will be tried. SNMP must be working to those chosen management addresses in order for Discovery to use them. Troubleshooting this further would require debugging to be enabled for the System discovery module. Then, the ngdiscovery.log would show the reachability checks.

That said, if SNMP is working to these addresses, but TFTP/telnet/SSH do not, an alternative woud be to use resolve by sysName, resolve by name, or none. None might be an acceptable option as that will use the address by which the device was discovered.

View solution in original post

Joe Clarke · ‎10-25-2010

What version of LMS are you using? LMS 3.2 should not use an unreachable address for management. If LMS determines that an IP is unreachable, it should skip it and move on to the next IP. If you're running LMS 3.1 or 3.0.1, make sure you apply the consolidated patch I talk about in the troubleshooting section of the document as this should add support for unreachable address detection.

KAROLY KOHEGYI · ‎10-25-2010

We use the newest version of LMS.

The problem that these ip Adresses from all VRF are reachable via ping for troubleshooting purpose.

The management protocols ( snmp, tftp etc ) are blocked by firewall's or NAC servers.

We had to introduce the VRF's becasue of NAC project. We found this problem during migration.

KAROLY KOHEGYI · ‎10-27-2010

Hi,

Since i have not received answer i think there is no solution for this problem.

So the lms discovery not able to distinguish VRF IP addresses. Every IP are same from this point of view.

Regards,

Joe Clarke · ‎10-27-2010

Sorry, I missed your reply. The code that checks for reachability uses SNMP to fetch the sysObjectID. If this fails, the next address in the ipAddrTable will be tried. SNMP must be working to those chosen management addresses in order for Discovery to use them. Troubleshooting this further would require debugging to be enabled for the System discovery module. Then, the ngdiscovery.log would show the reachability checks.

That said, if SNMP is working to these addresses, but TFTP/telnet/SSH do not, an alternative woud be to use resolve by sysName, resolve by name, or none. None might be an acceptable option as that will use the address by which the device was discovered.