Allan, > ... > > > Error: Accounting: logout: entry for NAS cnet-bsn port 1929379840 = > > > has wrong ID > > > I found the location of the message.... rlm_radutmp.c > > > > .. in my case, Shasta's port number won't change for each user. this = > > numer is compiled based on VCI/VPI/Slot/Module=20 > > numbers and it is not based on session ID that is the only unique = > > identifier of the subscriber. > > Well, THAT sucks. Then what the heck is the use of the NAS-Port? > Why is the NAS sending a numeric NAS-Port, which is constant for many > users? > > Unfortunately, the RFC does NOT require the NAS-Port to be unique > for each user. It DOES however state that the NAS-Port is meant to be > the physical port number on the NAS. > > So Shasta is using it in a way that's NOT expected by most RADIUS > servers. > > This also means that the Framed-IP-Address+ thing won't work, > either.
Shasta has an IP Pool configuration... in fact, shasta is much more complex that a simple RAS. It is an box that can create virtual routers at trunk side and virtual switchs at access side. I can send you many informations about this box if you want. > > > Alan, it is possible or fisible to make the rlm_radutmp module use the = > > sessionID in place of portnumber? > > It's already using the session ID. The error message comes out > because the saved session ID is NOT the same as the one in the packet. > The session ID is the same and is 8 character long. I think that the index of radutmp file is port number and when FR receives a STOP accounting packet, the session ID at port position is not the same because many others subscribers has logged in before and, logically, the session ID at that port number position is not the same of the STOP package. > This MAY be because the session ID is truncated in the 'radutmp' > file. The data structure currently only defines it to be 8 > characters. Maybe it needs to be more. > > If the session ID is longer than 8 characters, then we can increase > it's size in the data structure, and the error message will disappear. > > Alan DeKok. > > - > List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html > - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
