Re: WLC 2125 Rebooting

omarmontes · ‎11-21-2008

Hi guys, we are having some trouble with a WLC 2125 with only 6 APs, but like 40 clients trying to associate.

Version: 5.1.151.0

The issue here is that the WLC reboots every 6 or 8 hours for no apparent reason. CPU is always at 0% - 1% and mem usage is at 35% - 40%. I've heard in this forums that this kind of WLC is not recommended for more than 20 clients.. so I don't know if we are over using the box.

Maybe is an issue of software version or something, but in cisco.com there is only one version for this model of controller.

I'm adding an attachment with the crash file:

thanks!

jeff.kish · ‎11-21-2008

I would investigate a software upgrade or open a TAC case to see if this is a known bug. A 2125 can support 25 access points, so if 40 clients is pushing the limit, that's a major design flaw (just more than one client per AP??).

Otherwise, try tweaking things. Ask people to hardware in one day to reduce client count and see if it crashes. Disable AutoRF and see if that helps (Wireless Tab).

wesleyterry · ‎11-21-2008

Well, I see in the realease notes for 4.2.173.0 (October 31, 2008), there is a resolved caveat on the 2106 that causes a crash of bcastReceiveTask which is your crash. But you're not on a 2106 (though it is the same Product series).

With that said, there are no newer 5.1 releases, and I keep hearing not to go to 5.X yet.

With that said, assuming you have support, I would open a TAC case and demand they give you a fix for your crash. They may have a support fix that is just not available to the public.

Obviously they didnt fix the 2106 error untill October, so it would make since for the same error to be in an older release of software even though it is a higher software number...

wesleyterry · ‎11-21-2008

Oh and I may have a suggestion as a work-around. I once had a controller that after a few weeks would crash every 11 minutes. Had TAC send me a new one, reloaded the same config, and then 2 weeks later, every 11 minutes it would crash. I could reboot it all day long and it would always crash 11 minutes later.

What I found however, was that if I reset the controller to factory defaults, got it up and running again and then reapplied the configuration, I could get a few more weeks out of it.

So as a suggestion, get the config, reset to factory defaults and put the config back on it. See if that fixes it.

dennischolmes · ‎11-26-2008

Reboots usually occur to one of three reasons on a 2100 series controller. First is power. The power supply is very susceptible to power spikes and brownouts. If the power spikes, the controller will reboot. Second, memory leaks on the controller. When using Web Auth the controllers CPU and memory resources are consumed fairly quickly. If the code does not effectively release those resources the system will eventually crash. There will be a record of this in the Tech Support section under management on the controller. Third and lastly, thermal issues around the controller. This will show in the event logs as a warning that the controller has exceeded it temperature threshold. If it gets too hot it will shut down completely.

We are just completing the largest single deployment of 2100 series controllers in Cisco's history deploying of 200 2125s and a number of other 2100 series controllers and these are the major reboot issues we have seen.

omarmontes · ‎11-26-2008

Hi Dennis, thanks for your reply.

I'm gonna check the power supply issue. The temperature in the controller is around 38Â°C. And we are indeed using Web Authentication, but the CPU is almost at 0% and Mem usage is right now at 37%.

In the crash file, the crash mentions the task name of "bcastReceiveTask", but i couldnt find anything about running out or resources or something.

thanks again!