Hi, Marcos - we have an Aruba SE in our local area who took a look at this for us. He was able to correct it this way (quote from his e-mail to me) -
I added this to the 5400r configs…
snmp-server response-source dst-ip-of-request
Once I did that, and had Airwave poll again, they came back up.
Here’s my theory – did ANY thing else change in terms of firewall, ACL, etc. that might have caused this? Did you upgrade the 5400 firmware?
I don’t think this is an Airwave issue as my SNMP test client on my laptop also had the same issues… What happened is that the response was coming back through another IP interface on the 5400r and never making it back to Airwave. <end quote>
We now have snmp communication with our 5400s restored.
Just a couple of notes - in answer to the question, did we make any other changes, such as ACLs or firmware in the 5400s, the answer is no, we didn't - the only thing that changed was the AirWave version. Also, I thought that we still had SNMP to Solar Winds Orion during this problem, but further investigation showed I was wrong - the nodes showed as 'up' in Solar Winds because they were responding to ICMP requests, which Solar Winds uses to indicate up/down status; but Wireshark traces from the Solar Winds server to the switches showed that they were not responding to SNMP 'get' requests.
I wasn't aware of this, but evidently AirWave also uses ICMP to poll for status, but unlike Solar Winds, if there's no response to SNMP messages, Airwave marks the node as 'down'. So to my eye, it looked like Airwave had no SNMP, but Solar Winds did, when, in fact, neither of them did.
In any case, the configuration change to the 5400 switches that's referenced in the above quote fixed the problem - I'm still not sure what broke it in the first place, whether we had that config and it changed or if we never had it before and it wasn't necessary in the previous code version of AirWave - all that is speculation at this point.
Hope this may help you out.