Wireless Access

Reply
Occasional Contributor II

RAP Problem with 6.1.3.4/5

Hello all,

 

Has anyone any thoughts on something that is happening to us since upgrading from 6.1.3.1 to 6.1.3.4 (and continued in 6.1.3.5).

 

We have a master controller at the head office (3600).  We also have a few sites with local controllers.  We also have sites WITHOUT local controllers that have RAPs (in bridge mode) talking back to master controller.  Since at least 6.1.3.4, we have noticed the following:

 

At apparently random times, devices near a particular random AP will all lose connectivity (though the SSID's are still being broadcast).   These SSID's that lose connection are all WAP2 802.1x back to a windows RADIUS server, or a WPA2 AES passphrase.  Other AP's at that site are still fine and working...though we have had the odd time where it is more than 1 at one site.  

 

There does seem to be a link between usage and the problem...the heavier the wireless traffic, the more likely the problem is to happen to the loaded AP.

 

The only way to get the devices authenticating back through that particular AP is restart it...then everything is fine for a time...

 

We tested the same AP with an OPEN SSID (even applying the SSID at the time of the problem while the other SSID's had just failed), and even when the AP stops accepting authenticated connections, it still allows OPEN connections just fine.  The controller shows no errors with the AP at the time it happens.

 

If you manually assign an IP address etc to the client, you still are unable to get any connectivity through the AP on the encrpyted/authenticated SSID.

 

We have since dropped the controller back to 6.1.3.1, and we no longer see this issue happening.

 

Not sure what to look for...Aruba support has been great, but the problem occurring at random times, and the schools needing access has made getting traces extremely difficult....we have given up that solution...i.e. hoping the bug is known or gets fixed some other way....i.e. desperation fix.

 

 The sites with local controllers have not had this problem occur ever.  Seems to be related to RAP/bridge mode or both....

 

Open to any thoughts or suggestions....

 

I did notice another thread about some weird issues with password complexity making authentication fail and another user dropping back to 6.1.3.1 to solve that issue...related perhaps?

 

Grant

 

Occasional Contributor I

Re: RAP Problem with 6.1.3.4/5

Spoiler
 

Hello Grant,

 

For my part off view, this is a bug!

 

I did some tests and find out the following:

 

During upgrade process the ap puts the new code in a "second bank"

After the upgrade has been finished, the ap will reboot to load the new code.

The problem now is, that the ap always load the old code from the older bank.....

After that, he will do the upgrade again.......... never stops!!

 

If you connect the ap with serial-cable you have to perform following comand: "printenv"

inside the output you will find the following line: "os_partition = 1"

if this is 1, the ap will always boot from the old bank.  You should change this to "0"

 

use following command:  os_partion 0

                                               save

                                               boot

 

I hope this was clear for you. Otherwise i can give you some screen-dumps.

 

Greets,

 

Peter D'Affnay

 

 

 

 

The ap will upgrade the software on another bank inside the controller:

 

 

Meestal van de APs hadden hun 6.1.3.5 software kunnen opladen maar voor een onbekende reden blijft hem rebooten met 6.1.2.3 !

Dus met gevolg dat de controller “software mismatch” zag en de APs waren constant aan “upgrading” en “rebooting” te doen.

Met veel testen en troubleshooting heb ik gevonden dat een “os_partition” flag op de AP niet verandert wordt !

Via bootmode (serial cable) heb ik hem geforceerd op juiste partition te booten en het was in orde.

=> ik zal wel een case bij Alcatel openen.

Peter D'affnay
Guru Elite

Re: RAP Problem with 6.1.3.4/5


GrantT wrote:

Hello all,

 

Has anyone any thoughts on something that is happening to us since upgrading from 6.1.3.1 to 6.1.3.4 (and continued in 6.1.3.5).

 

We have a master controller at the head office (3600).  We also have a few sites with local controllers.  We also have sites WITHOUT local controllers that have RAPs (in bridge mode) talking back to master controller.  Since at least 6.1.3.4, we have noticed the following:

 

At apparently random times, devices near a particular random AP will all lose connectivity (though the SSID's are still being broadcast).   These SSID's that lose connection are all WAP2 802.1x back to a windows RADIUS server, or a WPA2 AES passphrase.  Other AP's at that site are still fine and working...though we have had the odd time where it is more than 1 at one site.  

 

There does seem to be a link between usage and the problem...the heavier the wireless traffic, the more likely the problem is to happen to the loaded AP.

 

The only way to get the devices authenticating back through that particular AP is restart it...then everything is fine for a time...

 

We tested the same AP with an OPEN SSID (even applying the SSID at the time of the problem while the other SSID's had just failed), and even when the AP stops accepting authenticated connections, it still allows OPEN connections just fine.  The controller shows no errors with the AP at the time it happens.

 

If you manually assign an IP address etc to the client, you still are unable to get any connectivity through the AP on the encrpyted/authenticated SSID.

 

We have since dropped the controller back to 6.1.3.1, and we no longer see this issue happening.

 

Not sure what to look for...Aruba support has been great, but the problem occurring at random times, and the schools needing access has made getting traces extremely difficult....we have given up that solution...i.e. hoping the bug is known or gets fixed some other way....i.e. desperation fix.

 

 The sites with local controllers have not had this problem occur ever.  Seems to be related to RAP/bridge mode or both....

 

Open to any thoughts or suggestions....

 

I did notice another thread about some weird issues with password complexity making authentication fail and another user dropping back to 6.1.3.1 to solve that issue...related perhaps?

 

Grant

 


Do you currently have Airwave?  Even if you don't, request and install an evaluation copy so that you can track historical performance as well as logging on your controller and access points.  That will allow you to assign events to logs to possibly performance.  It also might help you get to a solution to your issue with the help of support to interpret the data.

 

 

 



Colin Joseph
Aruba Customer Engineering

Looking for an Answer? Search the Community Knowledge Base Here: Community Knowledge Base

Search Airheads
cancel
Showing results for 
Search instead for 
Did you mean: