VISP Systems Administration <[EMAIL PROTECTED]> wrote:
> The infamous line:
> datestamp : Error: Dropping conflicting authentication packet from client
> clientname.realm.net:1645 - ID: xxx
Argh... I put a fix in the other day for non-threaded code. But
that may not help the threaded version.
> I am currently running the snapshot dated 08/14/01, and this issue seemed
> to crop up recently after perhaps this snapshot, or could have even been
> beta 2. The issue only seems to start up after the radius server has been
> running a couple hours and processed a few thousand requests. Then these
> conflicting lines begin showing up in the logs, usually 2-6 per unique
> login. The logins still seem to be making it through; however, these lines
> persist.
Weird..
> I have checked TS settings, and tried multiple TSs to ensure it is not a TS
> error. But in most instances the TS is set to re-send packet 10 seconds
> after IF it has not received a reply. These conflicting packet lines seem
> to appear mere seconds, after the request, and well before the TS sends a
> 2nd request.
That's even weirder. It appears, then, that the same request is
being seen multiple times. ??
> A restart of the radius process clears these errors immediately, until
> about 2 hours later where they start up again.
Yeah, killing it seems to help.
> I am assuming a memory issue is perhaps causing this phenomenon. Ideas?
I'm pretty much at a loss. If you have the time and interest in
tracking it down, that would help *enormously*.
That is, a simple statement like "the build from May 7 works, but
May 8 doesn't" would allow us to track the problem down. Right now, I
can't even reproduce it locally, so I have no clue what to be looking
for.
Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
