[Chilli] stresstest coova chilli

Peter Warasin peter at endian.com
Fri Jun 3 16:17:15 UTC 2011


Hi guys

On 03/06/11 15:56, Johan Meiring wrote:
> This is a radius problem and not coova chilli.
> This is quite normal.
> Radius packets are UDP and can easily get lost (especially with lots of
> clients)
> 
> You need to make your radius server able to handle lost accounting packets.

ah this sounds reasonable. thank you!
I first also thought about maybe an issue with our local radius, but
then we faced this also with different external radius servers.

Is radius really that unreliable?? This sounds dreadful to me. Or did we
simply configure it wrong?


> Lets say your interim accountint update is set to 5 minutes.
> Every time you receive an accounting packet, update a "last seen" field.
> 
> Run a cron job every 5 minutes that "closes" sessions (sets accounting
> stop time to something other than null) where the "last seen" field is
> older than 15 minutes.
> 
> Hope this makes sense.

Well, we do this so right now as a workaround (cron job). But i would
not consider this "the fix". Shouldn't chilli make sure accounting data
is correctly stored?
I'm a little perplex now :)

Well, thank you for your help! You save me a lot of time by not
searching on the wrong side.


peter

-- 
:: e n d i a n
:: open source - open minds

:: peter warasin
:: http://www.endian.com   :: peter at endian.com


More information about the Chilli mailing list