From: domenico d'a. <dom...@ya...> - 2011-04-23 11:05:05
|
Hello I tried to stress liquidsoap with 20 simultaneous processes. First 5 then 10 then 20 After a few minutes to process liquidsoap ceased (all twenty at the same time) giving me the error: 23/04/2011 11:55:13 [clock.wallclock_main: 2] Too much latency! Resetting active sources ... 23/04/2011 11:55:13 [pf_u120: 3] Closing connection ... 23/04/2011 11:55:13 [pf_u120: 3] Connecting mount for pf_u120 so...@ww... ... 23/04/2011 11:58:22 [pf_u120: 2] Connection failed: Could not read data from host! 23/04/2011 11:58:22 [threads: 3] Thread "wallclock_main" exited. 23/04/2011 11:58:22 [main: 3] Shutdown started! 23/04/2011 11:58:22 [main: 3] Waiting for threads to terminate ... 23/04/2011 11:58:22 [main: 3] Cleaning downloaded files ... 23/04/2011 11:58:22>>> END LOG There 's some parameter I can tweak to overcome this problem? Thanks Example file liq -------------------------------- #!/usr/bin/liquidsoap def my_request_function() = # Get the first line of my external process result = list.hd( get_process_lines("php /localhost/liqdynsingle.php 101 674")) # Create and return a request using this result log(result) request.create(result) end set("log.file.path","/var/audiopart/log/pf_u120.log") #set( "server.telnet", true ) set("log.stdout",true) radio = request.dynamic(my_request_function) radio = smart_crossfade(fade_out=1.5,fade_in=1.5,radio) radio = normalize( radio ) full = fallback(transitions=[def_trans,end_trans], track_sensitive=false, [radio, blank()]) output.icecast(%mp3, host = "localhost", port = 8000, password = "password", mount = "pf_u120", mksafe(full) ) |
From: David B. <dav...@gm...> - 2011-04-25 10:50:43
|
Hi, On Sat, Apr 23, 2011 at 1:04 PM, domenico d'agostino <dom...@ya...> wrote: > 23/04/2011 11:55:13 [clock.wallclock_main: 2] Too much latency! > Resetting active sources ... This is a normal consequence of stressing liquidsoap. More precisely, if it doesn't have enough computational resources (e.g. CPU) it'll lag, and eventually reset its outputs to avoid absurd latencies. This can be tweaked through the root.max_latency setting (which should maybe be reset to clock.max_latency, or made a parameter of clock(), but this is off-topic). > 23/04/2011 11:58:22 [pf_u120: 2] Connection failed: > Could not read data from host! I can't say much about this error. What is pf_u120? If it's an icecast output, it might be a little weird that it can't reconnect because of the stress. Cheers, -- David |