>> When I run a /etc/init.d/radiusd reload or restart, the first 20 minutes >> I get a lot of >> >> Fri Jan 21 10:33:51 2005 : Info: The maximum number of threads (32) are >> active, cannot spawn new thread to handle request
> It takes a bit of time to reload/restart the server, and during that > time, the clients continue to send requests. As a result, the server > may have a backlog of requests to process. > Still, taking 20 minutes to process the backlog is a bit much. >> I run a mysql in backend, but this server reports no significant load. > Is FreeRADIUS getting *fast* responses to its queries? If not, then > the delay is all in the DB. The whole problem seems to be related to the radutmp-file. Since I use sql in backend, I commented out the radutmp from the accounting-section (but kept the definition itself in radiusd.conf,because the radzap keeps looking for it) and now the speed and responsetime from the radius-server is definitively impeccable :) Because I use mysql, I had to do a few modifications to the radzap-program. Basically, I have commented out all "exit(1)" when the program fails to find the user, fails to verify the connection to the radius-server etc - and I always start radzap with -r <radiussrv-ip> NAS-IP port user This is obviously not a perfect solution, but it works. -- Med vennlig hilsen/Sincerely Alfred H. Dahl - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
