Handhelds disconnects randomly

[muratg7777]

Hi,

We have a 4404 WLC. Handhelds are disconnecting randomly. I am getting :

*Dec 17 14:32:02.151: %DOT1X-3-MAX_EAPOL_KEY_RETRANS: 1x_ptsm.c:407 Max EAPOL-key M3 retransmissions exceeded for client 00:0b:6b:af:1b:as

these logs. I set the EAPOL key timeout ve retries to maximum values but still getting the logs.

We use like 30 handhelds. 7-8 of them disconnects randomly. The handhelds are reconnecting after 15 minutes which is along time for us. I have looked at the handheld logs it retransmits the key and gets timeout. I have tested the field with 3 handhelds while roaming. None of them disconnected but sometimes handhelds disconnects. Maybe this happens on same handhelds. I will look for it.

I have searched the forum but couldn't find any solution. I use agressive load balancing. WPA/WPA PSK for authentication.

Software Version                 6.0.196.0

Any trick on handhelds or on WLC side that can be done?

I found this article = http://intermec.custhelp.com/app/answers/detail/a_id/9432

Will try to set these setting on the handhelds. Maybe some of you experience the problem.

[George Stefanick]

My suggest is to reverse engineer and get back to simple. By this I mean, lets turn off all the bells and whistles.I dont know your deployment but you mentioned agressive load balance is enabled. I would disable this. I would also disable "session timeout". You can find this @ CONTOLLER->WLAN->(select WLAN)->ADVANCE tab... Its a check box .. Uncheck this and test.

What more can you tell us about your controller config ?

[muratg7777]

Hi George;

The APs are 1310. We have one ssid. On the 802.11b/g settings, 1 mbps is Mandotory and all the other rates are supported. The other configuration is on default settings.

I did what you said. If agressive load balancing disabled, does this mean; the wlc won't load balance between APs?

[George Stefanick]

Aggressive load balance is client specific. The controller will try and manage clients on APs. Cisco recommends to disable this feature for voice clients. You dont want voice phones getting bounced around. The WLC will ignore association request from clients on certan APs to distribute the clients on APs.

By no means am i saying this is your issue. Troubleshooting is a art. I always like to get back to basics, get it working, reapply settings and see where it breaks.

You can also drop into the CLI and do a client debug on one of your device and capture what the WLC is seeing. This will be helpful as well.

[muratg7777]

Wireless network is for only handhelds so it was good for throughput but I will see what will happen.

Also intermec supports cckm, I will enable on handhelds.

I couldn't do debug on client side because it happens randomly. when I do site survey nothing happens but after a couple of hours or days the clients disconnects and not reconnecting again for 10-15 minutes time.

[bbxie]

Try to upgrade WLC to 7.0.98, 6.0.196 has some serious bugs related to random disconnection:

CSCtf34858 Client can't transmit traffic if it reassociates to an AP within 20 sec
CSCte89891 Radio may stop transmitting beacons periodically

[muratg7777]

I am still getting same error messages.

I heard that bugs but As far as I remember they were for 5500 Controller and there was a warning on cisco support when trying to download image for 5500.

4400 Controller use same code I think it affects too.

I searched the bug:

Fixed-In http://tools.cisco.com/Support/BugToolKit/images/Field%20Definitions.html
12.4(21a)JA2
12.4(21a)JY
7.0(94.114)
12.4(23c)JA
12.4(10b)JDE

It was modified last year. Does this 2 bugs fixed on 6.0.199.4(ED)

I couln't find info on release notes.  Which is recommended 7.0 or 6.0 (latest release)??

[bbxie]

These bugs should be related to both 5508 and 4400, pls. see Cisco's software advisory:

http://www.cisco.com/web/software/Wireless/Deferral/Software_Advisory_6_0_196_0.html

You can try both versions to see if the bugs be fixed. If your problem still not fix after upgrade, you have to open debug(debug client xxx) in WLC, use some wireless analyzing tools to capture the packets and check what happened during the disconnection(always remember to use NTP to sync all the related equipments' time). And of course, call Cisco TAC.

Before that, you definately should do some site survey to see if there's any coverage hole, also do some testing especially with simple environment as suggested by George. For example, firstly choose open, no ecryption for the WLAN, no session time out, no aironet extention, no dhcp(using static IP), enable low data rate such 1M/2M, etc. then test it, if problem disappear, then add more parameters towards your final enviroment to see which parameter cause the disconnection. If problem still exist, you need to check Intermec wlan driver, call TAC.