11-14-2013 04:39 AM
We have NNM 9.20(UNIX). Our internal client has insisted that SNMP queries from our NNMi management server has caused issues on specific nodes. So we put those nodes in the Excluded IP address section. However, based on the client's node logs(i.e.,community strings & ip address from the NNMi management server shows up in their logs), the SNMP queries continue. How can we have NNMI do the following? "Do not ping, do not SNMP query, do NOTHING on the internal client's nodes". Thanks in advance for any tips
11-14-2013 11:51 AM
I would ensure that those addresses are not covered by your discovery ranges. Excluded IP addresses behaves a little differently - if you discover a device, and it has some of those IPs, it will still discover the device, but act like those interfaces don't exist.
If those systems are truly so sensitive, they should be disabling SNMP on the devices themselves. Or at least using tight ACLs + different community strings. That is, if they are really serious about SNMP causing problems on those. Or maybe they're just looking to blame the NMS, because, well...it's always the NMS's fault.
11-14-2013 12:01 PM
Thanks, Lindsay. Unfortunately those IP addresses are covered by our discovery ranges. We have such a wide discovery range that we can't help but try to "discover" the Excluded IP addresses. I think the option will have to be as you suggested, shut off SNMP at the node level. A kludgy way to go for sure, but at least the internal client can't point fingers at the horrific SNMPWALK! :)
11-14-2013 12:22 PM
Or I suppose you could add the device to NNMi, but set it to Unmanaged? That might actually be the best solution here.
11-14-2013 12:26 PM
11-14-2013 11:42 PM
You said: "Do not ping, do not SNMP query, do NOTHING on the internal client's nodes".
I would simply change their management mode to either 'Not Managed' or 'Out of Service'.
There is a great online help section, search for "Understand the Effects of Setting the Management Mode"; you will find:
For nodes, setting the Management Mode to Not Managed or Out of Service has the following effects:
- No incidents are generated for the node
- The node's SNMP Agent is excluded from fault polling.
- The node's interfaces or addresses are excluded from fault and performance polling.
- NNMi quits gathering Node Component data.
- NNMi deletes all Polled Instances associated with the Not Managed or Out of Service node.
- The Active State for any Custom Poller Nodes associated with the Not Managed or Out of Service node becomes Inactive.
- The node is removed from any associated Router Redundancy Groups.
- Traps related to the node, interface, card, node component, or address, (for example, coldStart or warmStart) are not stored.
- The node is excluded from discovery.
- Actions ? Polling ? Configuration Poll is no longer available for this node.
- The status of a node is set to No Status.
- Actions ? Polling ? Status Poll is no longer available for the node or incident related to that node.
Is this what you want to achieve?
If you find that this or any post resolves your issue, please be sure to mark it as an accepted solution.
11-15-2013 03:38 AM
I am liking this option "Unmanaged" that you and Lindsay have proposed. The best solution is convincing the client that SNMP is not causing issues on his devices. Thanks.
11-15-2013 03:44 AM
Quick question here. If I set the node specific communications settings to use only ICMP, can the node still send back SNMP information on its own and it will be received by NNMi management server? Those nodes have the SNMP configuration set to have the NNMi management server as the trap receiver. The reason I ask is that we trouble ticket on SNMP information sent back by those nodes. Thanks.
11-17-2013 10:20 PM
11-17-2013 10:27 PM
For all your queries I would advise you to follow these steps even if it is bit lengthier to do,
1. Identify and list devices for which SNMP queries should not be sent from NNM
2. Remove these devices from NNM
3. Create SNMP profile for each device in NNM for discovery of the device as ICMP. Please make sure you would give
wrong community string. For creating In NNM console goto Configuration->Communication configuration->Specific
4. If needed create a new Nodegroup and add these devices to the node group
5. Identify and remove if any of the devices are still listed in any existing nodegroup which can use SNMP access. You can
use the command nnmnodegroup.ovpl for this ( Refer NNM reference guide for details of this command ).
11-18-2013 03:53 AM
I think this is the route to go. I can still ping the device, but then do no SNMP queries. Plus the device can still have SNMP running and send traps to the trap receiver. Thanks.