Mobility Express - clients disconnecting around every hour

[ktsaved]

New setup of 1852e access points in production area using 802.1x authentication to Windows NPS server, machine certificates.

Also have printers connecting with NPS user name and password.

Computer clients authenticate and connect fine. The problem is, viewing monitoring of clients, after about an hour to an hour and a half, they disconnect and re-authenticate. So the list is full of computers being connected for 30 seconds, 28 minutes, 1 hour 14 minutes, etc. Net sh wlan report shows the reason as "The network is disconnected because the user wants to establish a new connection".

After searching all settings and GPO settings for the Wifi profile and finding nothing wrong, checked one of these computers in the lab that is connected to a different Wifi access point. Same GPO, Radius profile, SSID, etc., just a different access point instead of one of the Ciscos. The client still has Unifi in the office areas. Check the controller and this computer has been connected for over 4 days.

So, it appears this is not initiated by the client after all.

I have disabled optimized roaming on 2.4 and 5 Ghz as I couldn't find any docs on changing the threshold level. 

This is a new installation of 8 1852e access points, have been running for a couple of months and have been in full production use about a week. Client is having issues with disconnects to their web app used in their production process, causing issues.

Happens on both 5 and 2.4 Ghz radios, and on the strongest signals (-48 dBm) along with the weakest ones (-68 dBm). The clients are stationary - they aren't roaming between access points.

Set 5 Ghz down to 6 mbs min data rates and 2.4 to 12 mbs.

One change is that after disabling optimized roaming the Zebra printers connecting via user name and password are staying connected for 4 and 5 hours now.

The computers are Lenovo m720q with Intel Dual Band Wireless-AC 8265 which are AC Wave 2.

Power settings are set to maximum performance for the adapters.

Hopefully someone can provide direction as it isn't too good that I recommend Cisco equipment to replace Unifi and the Unifi provides a more stable connection.

Running the latest firmware.

Thanks.

[JPavonM]

Using optimized roaming is not recommended to avoid stations from been disconnected.

Barcode scanners use to have sticky connections to APs but it's recommended to let the device to roam using their internal algorithm decision.

With regards of the disconnections you see every 90 minutes more or less, check session timeout in the WLAN profile and increase it. But don't use 0 value as there are some well-known connection stability issues with some devices. It will be better to use 1-production-day as the value (I'm currently using 54400 secs in most of the deployments).

You can check this value with the next CLI command:

(controller) >grep include "Timeout" "show wlan 1"

And you can change it with this command:

(controller) >config wlan disable 1
(controller) >config wlan session-timeout 1 54000
(controller) >config wlan enable 1

HTH
-Jesus
*** Please Rate Helpful Responses ***

[ktsaved]

Thanks for the response.

I checked the config and the session timeout is currently 86400 seconds which equals 24 hours.

Other timeouts are ok - sleep is 12 hours or 720 minutes. User Idle timeout is Disabled.

The Zebra printers have been connected now for 16 hours since I disabled Optimized roaming. But something is causing the computers to disconnect and re-connect around every hour, or every so often - is it random. As I look through the list I see Uptime of 46 seconds, 1 hour 7 minutes, 23 minutes, etc. for the computers.

We also have Android handheld terminals and they do roam, but one I see now that has stayed at a location has an uptime of 4 hours.

This seems to point to the computers, but like I said in my original post, I have an identical computer in the lab - everything is the same except it is connected to a Unifi access point and has an uptime now of 5 days with no interruptions. So to me this rules out the computer alone, unless there is some kind of communication going on that Cisco recognizes and Unifi does not. This part I need help with.

[tomi.hakkarainen]

Hi,

did you manage to get this issue sorted?

I have similar problems with Cisco mobility express and 5 1830i AP:s maybe shoukd consider changing to different solution would be ideal?

br, Tomi

[ktsaved]

In my case it was because of a GPO that I created.
The GPO was in place to connect clients automatically to the correct SSID. It ran every hour. In the client logs it showed as a disconnect when it went through the process of connecting.
The customer stated the clients were disconnecting randomly. When I saw this in the logs I thought this was the issue at first. The customer's issue was something different. It was due to the customer placing the stations in places behind metal machinery that was added that we had not expected when planning the network. It was solved with a combination of adding more WAPs and tuning the system for the best roaming/connecting to the correct WAP that we could get.