[ previous ] [ next ] [ threads ]
 From:  Peter Boosten <peter at boosten dot org>
 Cc:  m0n0wall dash dev at lists dot m0n0 dot ch
 Subject:  Re: [m0n0wall-dev] Suspected bug in captive portal (1.22)
 Date:  Wed, 11 Oct 2006 20:26:51 +0200
Jonathan De Graeve wrote:
>> Whenever the session ends by User-Request or Admin-Reset,
> AcctSessionTime
>> is updated correctly,
>> but whenever the session times out (AcctTerminateCause: Idle-Timeout),
>> AcctSessionTime holds an
>> incorrect value.
>> For instance: I had a session of 24 minutes (en seconds) but
>> AcctSessionTime would only register
>> 510 seconds. This didn't happen in 1.2.
> It is designed to work like this. In our opinion (and also the WISP
> draft) the user should only be accounted for the time actually being
> online.
> Assume this situation: users goes online at 00:00 and gets logged-off at
> 00:30. This would normally be a session-time of 30minutes. BUT if your
> idle-timeout is 10minutes and idle-timeout is the cause of the log-off
> the NAS will only send 20minutes as session-time. It doesn't take the
> hard-timeout,idle-timeout into account.

Ah, indeed, didn't think of this. Just did some calculations and you're
absolutely right.

It's even better this way.

>> Next I noticed that although I use interim updates, AcctSessionTime
> isn't
>> updated, unless the session
>> ends. This is also different from version 1.2.
> I know about this 'issue'. It will be fixed in a next version, no ETA
> (it already is in SVN) but I'm currently way to busy at my work with
> different stuff.

This is not a real problem for me, I just noticed it.

Thanks for your (very) quick response. I love the captive portal btw,
keep up the good work.