Controllerless Networks

Hello,

 

We have an access controller (developped in-house) to deliver captive portal which is in front of the IAP cluster.

 

From this access controller, I cannot snmpwalk directly the VC.

When I attempt to walk any of the other APs on the network it works, then I try walking the VC and it works.

After some time, snmpwalking the VC directly stops working.

 

The network is designed as such

 

Access controller -> Cisco Core Switch -> Aruba S2500 edge switches -> IAP

Everything is on the same subnet (not my design :) )

 

Any ideas?

 

No responses I guess.

I am also experiencing the following where it sometimes responds and sometimes does not.

If a VC would transfer over to another AP, the snmpwalk breaks.

IAP now version 3.4.0.1

hey pmonardo, I've got a potential bug filed for this within a client deployment. PM me and we can validate your setup as well as the suspected issue.

Sure thing.

Fix has been released in 3.4.0.3.

Not fixed in release 3.4.0.3 and neither in 4.0.0.1.
Still working on it

Do you have debug level enabled on the iAP's and if soo is there anything that happens everytime the VC stops reposding.

Do you have SNMP walk the VC IP and the AP IP at the same time and you are not getting replies from either?

I've done the following:
Taking a pcap at the uplink
Taking a pcap at the AP
Enabled SNMP debug on the IAP

Aruba has identified this as a bug and I am awaiting an answer

I told you what settings I applied on my iAP and I'm still able after 48 hours to get SNMP replies from my VC.

Tried them out and let us know :)

I tried the same with my IAP and it does not work.

Again, Aruba has filled a bug and was able to replicate it.

Do you have any update on this? I have the same problem, I can poll the the IAPs one-by-one, but not the VC IP address.

No update as of yet.
Which version did you test?

hello,

we are on 6.3.1.2-4.0.0.3_41784, all access points are IAP-225.

regards,
Mark Szabo

if i query an AP directly this is what i get is pretty unusable, most of the counters are showing 0 (SNR, Rate, Retries) even polling multiple times.

this has nothing to do with snmp, it's just one more thing not working, but i am really unhappy with this whole instant solution, nothing works as it should. we had serious packet loss, disconnects, we had to disable client match, 80 mhz channels, etc. its full of bugs and looks like we are beta testers.

 

# snmpwalk -M /root/mibs:/var/lib/mibs/iana:/var/lib/mibs/ietf -m ARUBA-MIB:AI-AP-MIB -v2c -c public 10.220.9.119 aiClientTable | grep '.....Y'
AI-AP-MIB::aiClientMACAddress.'.....Y' = STRING: b0:5:94:92:b7:59
AI-AP-MIB::aiClientWlanMACAddress.'.....Y' = STRING: 18:64:72:e3:b0:1
AI-AP-MIB::aiClientIPAddress.'.....Y' = IpAddress: 10.220.9.189
AI-AP-MIB::aiClientAPIPAddress.'.....Y' = IpAddress: 10.220.9.119
AI-AP-MIB::aiClientName.'.....Y' = STRING:
AI-AP-MIB::aiClientOperatingSystem.'.....Y' = STRING: Win 7
AI-AP-MIB::aiClientSNR.'.....Y' = INTEGER: 0
AI-AP-MIB::aiClientTxDataFrames.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientTxDataBytes.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientTxRetries.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientTxRate.'.....Y' = INTEGER: 0
AI-AP-MIB::aiClientRxDataFrames.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientRxDataBytes.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientRxRetries.'.....Y' = Counter32: 0
AI-AP-MIB::aiClientRxRate.'.....Y' = INTEGER: 0
AI-AP-MIB::aiClientUptime.'.....Y' = Timeticks: (0) 0:00:00.00

 

 

Have you opened cases with TAC to resolve your issues? Solution will only get better if TAC is aware so they can trigger proper personnel.

 

have you tried a simple SNMPWALK and see what you get?

snmpwalk -c <community> -v 2c <IAP IP>

I am using the same IAP version as well I haven't had those issues. Not with 225's though.

we have tac case open with the packet loss, connectivity problems, client match, etc issues. the snmp issue is new for me, we do not have tac case about that yet. i will proceed with opening a case. but as i understand you have problems with snmp as well, filled a bug, aruba could reproduce and you are waiting for the solution.

 

i can snmpwalk the 9 accesspoint we have, but not the virtual controller. the virtual controller ip address does not respond to snmp queries. 

 

if i snmpwalk the individual accesspoints i got information only about the specific access point. for example aiClientTable only contains the clients connected to the queried access point.

This issue is addressed in IAP 4.1 which is forthcoming shortly. Thanks to pmonardo for helping isolate the issue with Dev/Eng at Aruba.

 

Adam

I can officially confirm IAP 4.1 allows SNMPwalk to the VC