duplicate radacct entries

Ski Mountain ski_the_mountain at yahoo.com
Thu Feb 16 05:21:23 CET 2012


I did not modify the default schema.

I am just using the radius client that is in debian in order to authenicate pptp clients. Everything is on the same server so it is not a issue with packets being lost.  


i   radiusclient1                                                                 - /bin/login replacement which uses the RADIUS protocol for authentication               

I do blame the nas and have cursed at pptp quite a bit, not I just want the dam thing to only send one radacct packet.  





----- Original Message -----
From: Fajar A. Nugraha <list at fajar.net>
To: Ski Mountain <ski_the_mountain at yahoo.com>; FreeRadius users mailing list <freeradius-users at lists.freeradius.org>
Cc: 
Sent: Wednesday, February 15, 2012 10:48 PM
Subject: Re: duplicate radacct entries

On Thu, Feb 16, 2012 at 10:13 AM, Ski Mountain
<ski_the_mountain at yahoo.com> wrote:
> The problem appears to be that for the second entry there is never a acctstoptime entered into the databse,

Both entries have DIFFERENT acctsessionid (and thus, different
acctuniqueid). That means the NAS is sending acct for two different
sessions.

> or is there a way to make it so only entry is entered into the database.  where is just says 0000-00-00 00:00:00.

Blame the NAS.

It's possible that:
- on the first record, the NAS is only sending acct-stop
- an acct-stop packet is lost
- you modified the default scheme, since by default an accounting
record without acct-stop would have NULL in acctstoptime

I suggest you:

(1) log acct packets to detail file as well. e.g.

accounting {
...
detail
sql
...
}

that way you can examine the detail file later to see what packets FR
recieved from the NAS. That way you can correctly determine which of
the three possibilites above actually happen

(2) add an additional column: acct-status-type, and modify the sql
queries accordingly.
Similar to the above, but this way you can quickly determine whether
an acct-stop packet is lost.

-- 
Fajar





More information about the Freeradius-Users mailing list