Intel AX 201 lso is triggered

[William Gao 85157355]

Hi all,

We use 9800, version 17.3.3 and upgraded to 17.3.6 a month ago, we have AP 1702,2702,1832 and 9105 which is why we did not upgrade to a higher version like 17.9.x. Then we encountered the problem of WIFI disconnection at random, regardless of the AP model, regardless of the region completely random, users using the computer, for example, while writing an E-mail, wifi suddenly disconnected, in fact, wifi is still connected, but no Internet, about this point, Cisco's TAC explained that the idle timeout was set too long, so WLC did not kick the user even though there was no traffic at all. In this case, you cannot ping any device. In most cases, even the IP address is 169.254.x.x. Unless you restart your computer or reset your network card. Recently, however, this method has become increasingly ineffective, and rebooting still does not work, only connecting to a wired network.
Most of our users' laptops have a life span of less than 2 years, which are dell latitude and surface laptop. The network card of most computers is Intel AX201 160MHz.
We made a case to TAC, and they required us to capture packets. However, according to the result of packet capture, roaming occurred in the user's laptop, but in the process of re-establishing the connection, the user suddenly stopped responding to AP and there was no traffic. As for why the user stopped responding, they said they could not see.
Looking through the windows logs myself, I found that every time something went wrong there were a few fixed events: Lso is triggered, then event Roam Complete.
We tried to update the driver of wireless network adpater to 22.200.x, which is the newest, But not work.
We've been struggling with this issue for a month now, it seems like restarting wlc or turning AP off for a few days and then on for a few days, but still can't find the problem.
Please, please help. Thank you.

[marce1000]

  >....We use 9800, version 17.3.3 and upgraded to 17.3.6 a month ago, we have AP 1702,2702,1832 and 9105 which is why we did not upgrade to a higher version like 17.9.x.
                Note that 17.9.3 will support the (older) Wave 1 APs again , because of many customers still having them , perhaps it is worth trying , 

 M.

[William Gao 85157355]

Hi @marce1000 Sorry for the late reply, Thank you for the informaiton, it is very useful. We are making a new test, All 9105 AP connect to a new 9800 WLC, which is 17.9.3. And see if it still drop. If there is no problem, We will let 1 2702 to join the new WLC, if still no problem, then it should be the problem of WLC version. 

[shiwa]

Hi @William Gao 85157355 ,

We also have the same random user disconnection issue in our environment. We are using 9800 with version 17.6.4 on WLC and 9105/9124/9115 AP models. We are using Flex Connect mode. How is it going so far with TAC? Any progress on this? Appreciate your update. Thanks

[Bro William]

Hi Shiwa,

Sorry I didn't see your message, How is your WIFI now? Please see my reply at 11-19-2023 10:19 PM.

[JPavonM]

Are you using WPA3 or PMF? If so, try downgrading the WLAN security to WPA2 w/o PMF first, then if working fine upgrade to PMF optional, then to WAP3 Transition. This could be the problem, but mybe others like 'Required DHCP' enabled that I recommend to disable, or even increasing EAP timeouts like this:

wireless security dot1x eapol-key retries 4
wireless security dot1x eapol-key timeout 2500
wireless security dot1x group-key interval 54000
wireless security dot1x identity-request retries 4
wireless security dot1x identity-request timeout 20
no wireless security dot1x max-login-ignore-identity-response
wireless security dot1x request retries 6
wireless security dot1x request timeout 30

[William Gao 85157355]

Hi @JPavonM The problem SSID use different authentication method, one ssid is WLC MAC list, others are WPA+WPA2. We will contact TAC to talk about the suggestion you gave. Thank you very much!

[Klyle]

Did you ever get an answer on this? I'm encountering the same issue with 9105's.

[Bro William]

Hi Klyle，

Sorry for reply late, I finally have a clue.

Please try to update windows system, install all patches, And if still doesn't work, try download newest version of WIFI and Bluetooth driver from the device brand official website, not INTEL original version! 

We got some people whose wifi was fine, but after upgrading the INTEL orignal version driver, it went wrong, And became good again after roll back to old driver. The INTEL community admin said don't use their orignal driver, Try the custom driver from the official website of your device brand, Because we all surface, so we just install windows patches, We got 2 surface recovered.

Cause of the problem, CISCO TAC said, According to the packet capture result, the client does not reply to the AP during the handshake, To be exact, Nr. M3. They concluded that the client tried to roam but failed to establish a connection to the new AP. But they cannot explain why the client wants to roam, even the laptop is just on the desk.

We finally find it is problem after changing AP from 2702 to 9105. Same as you. We tried 9120 and 9164, they both work, 2702 actually is also work! My theory is 2702 was the high end or middle end at its time, just like 9120 and 9164 are now, some chips are better than lower model, So even an old high end perform good than a new generation low end. 

Fortunately, we were in time for the large-scale AP swap, so we took the opportunity to switch to 9164, Then most laptops are happy, But not all, Then we met a vendor, they said they met this situation before at siemens factory, It was very serious problem, hundreds of laptops cannot connect to wifi. they asked INTEL to made a specific driver for the factory finally. But for us, lucky things is fix the problem by update Windows system...

Let me know if it works or not! Good luck.