From: Romain B. <to...@ra...> - 2011-06-21 22:33:15
|
Thank Jeff, I did not know that.. I have run some tests regarding output.harbor and the load. It triggered a bug that is now fixed. All in all, the performances are not ridiculous at all. I would say that you could seriously consider using output.harbor for less than 500 simultaneous listeners and could think about using it with more caution with less than 1,000 simultaneous listeners. With more than 1,000 listeners, I am not sure about the behaviour of the code (yet!). Romain 2011/6/21 Jean-Francois Mauguit <je...@ra...>: > Hello, > > You can do the listener tracking from the icecast logs > > 86.73.225.61 - - [21/Jun/2011:17:09:18 +0200] "GET /Kaina-Radio HTTP/1.1" 200 112155525 "http://www.kainaradio.com/radio/Webradio.php" "Mozilla/5.0 (Windows NT 5.1) AppleWebKit/534.30 (KHTML, like Gecko) Chrome/12.0.742.100 Safari/534.30" 7018 > > IP > Date > Mountpoint > Data > Useragent > Duration (in seconds) > > Date - duration => connection time > Date => disconnection time > > HTH > > Jef > > On 21 Jun 2011, at 16:58, Romain Beauxis wrote: > >> I agree with david. However, the code has been written carefully in >> the hope that it would be efficient. It needs to be tested with a high >> load though :-) >> One interesting feature of output.harbor compared to icecast is that >> you can keep track of the exact moment when a listener connected and >> disconnected using on_connect/on_disconnect, thus tracking down the >> exact listening time. > > > |