On Fri, 31 Oct 2003, Jonathan Ruano wrote: > Hi there: > > I messed with rlm_ippool sources in the past but I decided to give the 0.9.2 > version a try, as I hadn't touched anything since July. > > Not doing stress, full load tests, but took a look at the sources and > remembered how stale entries are found and fixed. > > The ippool array is indexed by nas/port, so if we're to assign an IP address > to a dialup user using the same nas/port combination than a previous > (currently marked as active) one, then the latter must be a stale entry. > > This works great (any real-life experiencies to share, anyone?) for just one > ippool instance, but not when there are several.
Why? However many instances you may have they will all check for a stale entry for that nas/port combination in their respective databases. So where exactly do you see a problem? > I'll see if I can merge my > hacked version and 0.9.2's. > > Jonathan Ruano > > > - > List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html > -- Kostas Kalevras Network Operations Center [EMAIL PROTECTED] National Technical University of Athens, Greece Work Phone: +30 210 7721861 'Go back to the shadow' Gandalf - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
