Re: [Savonet-users] Read error Unix.Unix_error(56, "read", "") ???
Hi Romain >Ok, I am looking at your issue now. > >From what I see on the log, it seems like everything hapened has expected. > >The exception you saw on the log is generated because the reading on the >socket failed, and as expected you see that the harbor source disconnects him. > >Then, the second request that you see on the logs is in fact a metadata update > >request. It means that the source client tried to do a metadata update >request, which will suceed even if the audio feeding has failed. > >The thing that I don't see in the logs is the second source client connection >on the harbor. Perhaps it should need to look later in the logs ? > >So, my understanding is: > * The connection with the DJ was lost at some point > * The behaviour after that was not what you expected > >Hence, as such I cannot reallt judge whether the problem after the >deconnection comes from the source client (VirtualDJ) or from liquidsoap. > >It would be nice if you could send us more logs, in particular a bit before >and after the issue so that we can see the harbor source client connections. The summary of the problem is that for whatever reason (Unix error), the audio from the DJ stopped playing ont he radio (i.e. he got disconnected so it fellback to playing playlist music) BUT it did not run the on_disconnect. It always runs on_disconnect when the DJ disconnects or I kick him via telnet so for some reason with this unix error, it didn't run on_disconnect Here is a copy of the log with more info. Let me know if you need more: http://dev.plurlife.com/temp/log.txt Gabe-- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf___ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
Re: [Savonet-users] Read error Unix.Unix_error(56, "read", "") ???
Hi Gabriel ! Le dimanche 20 septembre 2009 21:20:08, Gabriel Afana a écrit : > So today I had something unusual happen for the first time. A DJ was > playing live (input.harbor) and about 40 minutes into his show, liquidsoap > fell back to the playlist and continued playing the normal playlist that > plays when there isn't a live show. > > The really strange part is I checked to logs thinking to see the DJ simply > lost his connection, however this wasn't the case. The logs showed that > he reconnected again and my on_connect and on_disconnect function logs > show that he did connect again but never disconnected (i.e. he was still > streaming). However liquidsoap was still playing playlist songs and > wasn't going to the input.harbor source. > > This has never happened before. I had to go in and manually run the > on_disconnect function since it never ran. I have a crontab running every > minute that checks and disconnects the DJ the minute his timewinow is up > (using telnet connection). Even if all else fails, this will disconnect > his input.harbor connection and of course should run the on_disconnect > function. The crontab ran and connected via telnet but the on_disconnect > function never ran (leading me to believe he was never connected anyways > because the radio was streaming the playlist and not his live > show). However, he connected but never disconnected (but was no longer > connected!!) haha confusing! I noticed an unusual error i've never > seen. Maybe this is what caused this strange occurance? Ok, I am looking at your issue now. >From what I see on the log, it seems like everything hapened has expected. The exception you saw on the log is generated because the reading on the socket failed, and as expected you see that the harbor source disconnects him. Then, the second request that you see on the logs is in fact a metadata update request. It means that the source client tried to do a metadata update request, which will suceed even if the audio feeding has failed. The thing that I don't see in the logs is the second source client connection on the harbor. Perhaps it should need to look later in the logs ? So, my understanding is: * The connection with the DJ was lost at some point * The behaviour after that was not what you expected Hence, as such I cannot reallt judge whether the problem after the deconnection comes from the source client (VirtualDJ) or from liquidsoap. It would be nice if you could send us more logs, in particular a bit before and after the issue so that we can see the harbor source client connections. Romain -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf ___ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
Re: [Savonet-users] Liquidsoap runs on noise.
Hi, Thanks for your reply! > Hi ! > > Le samedi 26 septembre 2009 08:48:33, Mathijs Vos a écrit : > > The stream is sent, however it contains nothing but silence. > > However during some experimenting, I decided to change the "blank()" > > fallback thing into "noise()". I rebooted my server and started listening > > the stream again. After about two seconds of noise, music started! Just to > > make sure it was the blank() to noise() that made the stream working > > again, I changed it back, rebooted and... it didn't work anymore. Then > > again I used noise() again and Liquidsoap was back in action. > > I don't really understand why it works with noise, but I can tell you why it > does not work with blank. > > The fallback operator by default selects a new source only when the track of > the current source ends. Hence, when you start it with a playlist and a blank > source, if the playlist is not imediately ready, which is very likely to be > the case everytime, it selects the blank source. > > However, since the blank source is a single track, it never switches back to > the playlist. > > In order to change that you should add the track_sensitive=false option to > the > fallback operator. This is for instance the way mksafe is defined: > > def mksafe(s) > fallback(id="mksafe", track_sensitive=false, [s, blank(id="safe_blank")]) > end > > > Romain I see, that indeed was quite a stupid mistake from my side. I'll test this tomorrow, although I guess it'll work. Thanks once again! Cheers, Mathijs _ Krijg updates van je vrienden en blijf op de hoogte http://home.live.com-- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf___ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
[Savonet-users] Liquidsoap runs on noise.
Hi there once again,
This time I've got a really weird problem with Liquidsoap.
The script I'm using is:
set("init.daemon.pidfile.path","/etc/liquidsoap/pidfile/pid.pid")
def replaygain_protocol(arg,delay)
x=get_process_lines("/etc/liquidsoap/replaygain.pl #{quote(arg)}")
if list.hd(x) != "" then
["annotate:replay_gain=\"#{list.hd(x)}\":#{arg}"]
else
[arg]
end
end
add_protocol("replay_gain", replaygain_protocol)
muz = fallback([
playlist(
default_duration=30., length=10.,
conservative=true, timeout=20.,
mime_type="audio/x-mpegurl",
"/etc/liquidsoap/muziek/afspeel.txt"
),
blank()
])
muz = amplify(1.,override="replay_gain", muz)
output.icecast.vorbis(
host="localhost", port=80,
password="hackme", mount="muziek.ogg",
stereo=false, quality=0.,
description="Testing the best thing",
name="Just_some_testingchannel",
muz
)
This script does somehow not work.
The corresponding part of the logfile is:
2009/09/26 15:37:48 >>> LOG START
2009/09/26 15:37:48 [protocols.external:3] Didn't find "ufetch"
2009/09/26 15:37:48 [protocols.external:3] Found "/usr/bin/wget"
2009/09/26 15:37:48 [main:3] Liquidsoap 0.9.1
2009/09/26 15:37:48 [lang:3] flac binary not found: flac decoder disabled.
2009/09/26 15:37:48 [lang:3] metaflac binary not found: flac metadata resolver
disabled.
2009/09/26 15:37:48 [lang:3] faad binary not found: faad decoder disabled.
2009/09/26 15:37:49 [threads:3] Created thread "non-blocking queue #1".
2009/09/26 15:37:49 [threads:3] Created thread "generic queue #1".
2009/09/26 15:37:49 [root:3] Waking up active nodes...
2009/09/26 15:37:49 [afspeel(dot)txt:3] Loading playlist...
2009/09/26 15:37:49 [threads:3] Created thread "root" (1 total).
2009/09/26 15:37:49 [afspeel(dot)txt:3] Playlist treated as format
audio/x-mpegurl
2009/09/26 15:37:49 [afspeel(dot)txt:3] Successfully loaded a playlist of 10
tracks.
2009/09/26 15:37:49 [muziek(dot)ogg:3] Connecting mount muziek.ogg for
sou...@localhost...
2009/09/26 15:37:49 [muziek(dot)ogg:3] Connection setup was successful.
2009/09/26 15:37:49 [root:3] Broadcast starts up!
2009/09/26 15:37:48 [src_4658:3] Switch to src_4657.
2009/09/26 15:37:49 [stderr:3] Tags present; no files processed
2009/09/26 15:37:49 [decoder:3] Decoder OGG chosed for
"/etc/liquidsoap/muziek/4.ogg"
The stream is sent, however it contains nothing but silence.
However during some experimenting, I decided to change the "blank()" fallback
thing into "noise()".
I rebooted my server and started listening the stream again. After about two
seconds of noise, music started!
Just to make sure it was the blank() to noise() that made the stream working
again, I changed it back, rebooted and... it didn't work anymore.
Then again I used noise() again and Liquidsoap was back in action.
For the working stream, this was written in the logfile:
2009/09/26 15:34:21 >>> LOG START
2009/09/26 15:34:20 [protocols.external:3] Didn't find "ufetch"
2009/09/26 15:34:20 [protocols.external:3] Found "/usr/bin/wget"
2009/09/26 15:34:21 [main:3] Liquidsoap 0.9.1
2009/09/26 15:34:21 [lang:3] flac binary not found: flac decoder disabled.
2009/09/26 15:34:21 [lang:3] metaflac binary not found: flac metadata resolver
disabled.
2009/09/26 15:34:21 [lang:3] faad binary not found: faad decoder disabled.
2009/09/26 15:34:21 [threads:3] Created thread "non-blocking queue #1".
2009/09/26 15:34:21 [threads:3] Created thread "generic queue #1".
2009/09/26 15:34:21 [root:3] Waking up active nodes...
2009/09/26 15:34:21 [afspeel(dot)txt:3] Loading playlist...
2009/09/26 15:34:21 [threads:3] Created thread "root" (1 total).
2009/09/26 15:34:21 [afspeel(dot)txt:3] Playlist treated as format
audio/x-mpegurl
2009/09/26 15:34:21 [afspeel(dot)txt:3] Successfully loaded a playlist of 10
tracks.
2009/09/26 15:34:21 [muziek(dot)ogg:3] Connecting mount muziek.ogg for
sou...@localhost...
2009/09/26 15:34:21 [muziek(dot)ogg:3] Connection setup was successful.
2009/09/26 15:34:21 [root:3] Broadcast starts up!
2009/09/26 15:34:21 [src_4658:3] Switch to src_4657.
2009/09/26 15:34:22 [stderr:3] Tags present; no files processed
2009/09/26 15:34:23 [decoder:3] Decoder OGG chosed for
"/etc/liquidsoap/muziek/5.ogg"
2009/09/26 15:34:24 [afspeel(dot)txt:3] Prepared "/etc/liquidsoap/muziek/5.ogg"
-- RID 1
2009/09/26 15:34:24 [src_4658:3] Switch to afspeel(dot)txt with forgetful
transition.
2009/09/26 15:34:27 [src_4660:3] Overriding amplification: 0.328095.
2009/09/26 15:34:27 [root:2] We must catchup 4.04 seconds!
2009/09/26 15:34:28 [stderr:3] Tags present; no files processed
2009/09/26 15:34:28 [root:2] We must catchup 3.74 seconds!
2009/09/26 15:34:28 [decoder:3] Decoder OGG chosed for
"/etc/liquidsoap/muziek/8.ogg"
2009/09/26 15:34:29 [root:2] We must catchup 2.99 seconds!
To me this seems like a really strange bug. Is there a way how I can solve it?
Or if it's really a bug, can you guys solve it for the next bugfix release or
Re: [Savonet-users] Liquidsoap runs on noise.
Hi ! Le samedi 26 septembre 2009 08:48:33, Mathijs Vos a écrit : > The stream is sent, however it contains nothing but silence. > However during some experimenting, I decided to change the "blank()" > fallback thing into "noise()". I rebooted my server and started listening > the stream again. After about two seconds of noise, music started! Just to > make sure it was the blank() to noise() that made the stream working > again, I changed it back, rebooted and... it didn't work anymore. Then > again I used noise() again and Liquidsoap was back in action. I don't really understand why it works with noise, but I can tell you why it does not work with blank. The fallback operator by default selects a new source only when the track of the current source ends. Hence, when you start it with a playlist and a blank source, if the playlist is not imediately ready, which is very likely to be the case everytime, it selects the blank source. However, since the blank source is a single track, it never switches back to the playlist. In order to change that you should add the track_sensitive=false option to the fallback operator. This is for instance the way mksafe is defined: def mksafe(s) fallback(id="mksafe", track_sensitive=false, [s, blank(id="safe_blank")]) end Romain -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf ___ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
