0 Replies Latest reply on Nov 26, 2009 4:40 AM by GaryP

    Bind Polling Engine to Specific Interface

    GaryP

      Hi all, I recently brought up a second physical NIC (nic_b) on the machine running my Orion NPM instance (9.5.0) with an address on a different IP network range and VLAN (network_b). Due to ACLs in place across my network I only allow SNMP access to many devices from the initial NIC (nic_a on network_a). Now Orion seems to be trying to use nic_b to send out some SNMP queries (there are some devices it tries to query that are actually on network_b but they still need to come from nic_a on network_a).

      How do I force the polling engine to only use nic_a to send and receive it's SNMP traffic (and ICMP, too, for that matter)?

      I'm running Orion NPM 9.5.0 on Windows Server 2008 Enterprise R2.

      *edit*

      Okay, I seem to have fixed/bodged this myself but I'm still curious to hear others' views on how to do this.

      nic_b was brought up purely to receive syslog and snmp traps so I set a very high interface metric on nic_b (100) and much lower (1) on nic_a. This seems to have had the effect of routing all traffic out through nic_a now (both interfaces have default gateways set, so can get anywhere on our network from a routing point of view).

      As I say, please shout out if you have a more elegant solution to this.