Controllerless Networks

I have some issues reported by a customer that when they move locations they are unable to reconnect and/or get an ip address.

29 x IAP215

6.4.2.6 - 4.1.1.7

I have checked a couple of times when it was reported and there were 2-3 APs where none of the clients had a valid ip address and only 169.254.  Each occassion was different APs.  Rebooting the AP seem to resolve it and any client connecting to it gets a proper ip address.

At one point during the issue I was able to do a pkt-dump on dhcp and what I could see was the client requesting an ip address and the server responding.  Unfortunately I could not verify the response was going back to the device over the wifi.

• port config on all aps checked and fine.
• helper address setup on switch
• address pool is nowhere near exhaustion.
• dhcp server is receiving the requests and responding

I will raise a TAC case for this as it is pointing to something up with the APs, but wondering if anyone has had similar issues.

What kind of clients are these?  Are they using encryption?  802.1x?

All device types as far as I can tell.

Dot1x and WPA2-AES.

Strangely when the issue occurs, if I reboot the particular AP then clients connecting to it will get the correct ip address.

Working on it with TAC.

I would suggest to check the logs of the switch too.

At the same time try connecting to the AP via SSH and check if you can reach the VLAN.

enable debug on a few clients and see the logs.

I have disabled 802.11r on this ssid about a week ago and so far have not seen the issue occur.

Having said that, I have no evedence whatsoever that this was actually causing the problems, but nevertheless that is what I'm seeing.

[EDIT] Clearly seems like some clients did not like dot11r enabled.  Unfortunately the customer did not supply the make, model and driver version of the affected devices at the time, so difficult to pin down.

Hi guys

I'm facing the same issue with IAP 205 but the 802.11r option is disabled here. If I reboot the VC the client become to be able to get IP again.

I'll open a ticket with TAC.

Luiz

I saw a bug ID fixed in 4.1.1.8 that looks like it could be what you´re experiencing. We have also seen the same issue on IAP215 but havn´t upgraded yet, it´s not very frequent.

We are having this issue as well. I restarted the master and everything started working again.

We do not use DHCP helpers and just push the machines onto the vlan.

We are on: 4.3.1-4-2.0.0_51112

I have not had it occur recently, however there may be another issue causing it.

Across the road from the site is a large radio station transmitter.  This is putting interference into the wires to the point that you pick up some ip phones and it plays the radio station.  The customer is testing the cable runs to these affected aps to see if there is any spurious signals in the wires.

We are having same issue with multiple location. First time I had ignored this issue, but after 1 month its impacted with multiple location with multiple users. I have open TAC for solution they suggested to upgrade OS Version from 6.4.2.6-4.1.1.11  to 6.4.2.6-4.1.3.1 But still having issue with IAP. I have restarted one AP issue resolved for that AP after half hour another access point behaving like same, I have restarted that Ap's also issue resolved. after restart all APs issue resolved. But next day same issue open. Still there has no solution found from TAC. 802.11r option has disable at all SSID.

IAP 225 Model.

OS version  6.4.2.6-4.1.3.1

I had a similar problem some time ago.

The solution here was to disable the DPI (deep packet inspection) feature on the IAP as this caused a memory exhaustion every ~60 days on some cluster members. Those members then allowed clients to associate in wireless but no network traffic was forwarded.

HI LongIsland,

Thanks for your support. I am not find DPI setting in IAP cluster. can you narrow down or show me step to disable DPI.

I checked with above command but commant not validate. no firewall dpi.

My IAP Cluster version 6.4.2.6-4.1.3.1_55246.

Hi Sumitchavan27,

see this post, where DPI is explained in context:

http://community.arubanetworks.com/t5/Aruba-Instant-Cloud-Wi-Fi/AppRF-with-Instant/td-p/187750

HI,

Thanks for your suggestion

I have run no dpi --> then Commit Apply command on Master IAP but still shows apps details.How to stop this with command.

---

@sumitchavan27 wrote:

HI,

Thanks for your suggestion

 

I have run no dpi --> then Commit Apply command on Master IAP but still shows apps details.How to stop this with command.

---

Think those details are historical data it collected before you turned off DPI. 

What is providing dhcp to your clients?

My DHCP server propered shared IP address but access point not shared that detaiils with user. In one cluster 2 or 3 access point where infected. If we restart infected(Hang) Ap then another ap will be hang. If we restart to master then issue resolved for a month.

Do your wireless and wired devices share the same dhcp server?

Yes same DHCP server, Location wise we have seperate DHCP server. but different scope for wired & wireless client.

Did tac advise you to get the tech support of the AP when this happens?

Yes Tac asked me to shared tech support logs when issue happend. I shared Tech support log also. They required DHCP Uplink capture and user PCAP logs for troubleshoot. I have already shared same details with TAC. but not yet solve this issue.