On Wed, Mar 08, 2000 at 04:38:51AM +0000, tmercado wrote:
> Hello Hugh,
>
> Ok, I'm running Radiator with a Trace 4. We have to wait to see what happend,
> anyway, I think that the delay is not the problem, because the DupInterval is
> seted to 60 seconds, so if the MAX TNT send a duplicate after a timeout
> occurs (i.e. 7 seconds), radiator must ignore it.
>
> The question now is, why is not happening that?
> So, to check for duplicates packets, Radiator compares all data in the packet?
> or the acct-session-id and NAS-IP-address fields only?
>
> The database is ok, I was testing it with radpwtst and I can do between 120 to
> 130 request trougth radiator in a second (i.e. authenticate a user and save start
> and stop packets for that user), so is very fast. At this time, the rate for
> the real system is between 3 to 5 request in a second.
>
It compares the whole packet (well the MD5 checksum from memory).
*hops on hobby horse*
This means if your duplicate packets aren't identical then Radiator can't
detect and ignore duplicates (eg Cisco has Acct-Delay-Time which changes
in value for each retransmitted packet).
Check your Trace 4 log and compare the Stop records and see what the difference
is.
*heads to sunset*
[EMAIL PROTECTED]
===
Archive at http://www.starport.net/~radiator/
To unsubscribe, email '[EMAIL PROTECTED]' with
'unsubscribe radiator' in the body of the message.
