Icecast - May 2018 - Why "has fallen too far behind"?

Hello,
  
 I need to get some values passed in the querystring request in the log 
file.
  
 Something like that bellow:
  
 ie: "GET /radio?id=1 HTTP/1.1" instead of only "GET /radio
HTTP/1.1"
  
 Do I need to change something in the source and compile? Or there is 
another way?
  
  
 Best,
 Thiago

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180506/a052b506/attachment.html>

Dear Mr./Ms. subscription,


On Sun, 2018-05-06 at 08:35 -0300, subscription at nextdial.com.br
wrote:
>  I need to get some values passed in the querystring request in the log 
> file.
>   
>  Something like that bellow:
>   
>  ie: "GET /radio?id=1 HTTP/1.1" instead of only "GET /radio
HTTP/1.1"
>   
>  Do I need to change something in the source and compile? Or there is 
> another way?
What exactly is your goal? Maybe you can give us a bit of an
understanding of the context.

Query parameters are to be interpreted by the server, in this case
Icecast, not external software. Therefore I suspect that there is a
better solution for your problem.

With best regards,

-- 
Philipp.
 (Rah of PH2)

Dear Philipp,
  
 Thanks for you reply.
  
 My goal is to show in our analytics page segmented data (official apps, 
partner apps, third-party apps, demographic and geo). To do so each app 
pass some values in the query string (ie: app id, user id, lat, lng). And I 
need to know those values in order to do that.
  
 Today we are using nginx as a reverse proxy to Icecast to get those data 
but it double the cpu/mem usage =(
  
 I am welcome to any idea. =)
  
  
 Best,
 Thiago
  
  

----------------------------------------
 De: "Philipp Schafft" <lion at lion.leolix.org>
Enviado: domingo, 6 de maio de 2018 09:30
Para: subscription at nextdial.com.br, "Icecast streaming server user 
discussions" <icecast at xiph.org>
Assunto: Re: [Icecast] How to log querystring values?   
Dear Mr./Ms. subscription,

On Sun, 2018-05-06 at 08:35 -0300, subscription at nextdial.com.br
wrote:
> I need to get some values passed in the querystring request in the log
> file.
>
> Something like that bellow:
>
> ie: "GET /radio?id=1 HTTP/1.1" instead of only "GET /radio
HTTP/1.1"
>
> Do I need to change something in the source and compile? Or there is
> another way?
What exactly is your goal? Maybe you can give us a bit of an
understanding of the context.

Query parameters are to be interpreted by the server, in this case
Icecast, not external software. Therefore I suspect that there is a
better solution for your problem.

With best regards,

--
Philipp.
(Rah of PH2)
 

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180506/afeef368/attachment.html>

Hello,
   
 After removing a mount from the xml config file and reloading the Icecast 
with the bellow command why it don't close the current mount connection?
  
 /etc/init.d/icecast2 reload
  
 There is another way?
  
 Best,
 Thiago


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180514/ce9a6ccd/attachment.html>

Well I am only guessing here, but I suspect if icecast works like other daemons
reload just reloads the config but doesn't terminate current connections.
You probably want /etc/init.d/icecast restart


On 15 May 2018 12:35:39 AM NZST, "subscription at nextdial.com.br"
<subscription at nextdial.com.br> wrote:
>Hello,
>   
>After removing a mount from the xml config file and reloading the
>Icecast 
>with the bellow command why it don't close the current mount
>connection?
>  
> /etc/init.d/icecast2 reload
>  
> There is another way?
>  
> Best,
> Thiago
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180515/a9473dd4/attachment.html>

Hello,
  
 We are seeing in Firebase this kind of errors:
  
     
com.google.android.exoplayer2.upstream.HttpDataSource$InvalidResponseCodeExc
eption: Response code: 401
  
 So, after 12h trying to reproduce the error we saw what's happening. If 
the listener connection buffer due to anything, Icecast remove it's 
connection and log the item bellow:
  
     [2018-05-22  23:17:24] INFO source/send_to_listener Client 220977 
(xxx.xxx.xxx.xxx) has fallen too far behind, removing
  
 My question is:
  	Why Icecast need to do this? 	Why the client can't reconnect? 	The error 
401 is when the client tries to reconnect and our auth deny? 
    
 Best,
 Thiago


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180522/6fbe411d/attachment.html>

Hello,
   
 We are using Icecast in a 2 vCPU VPS running Ubuntu 16 LTS.
  
 In our lastest test we saw in HTOP that Icecast was consuming about 94% in 
one CPU and, in the other one, ~57%. We had ~2400 simultaneous listeners 
96Kbps AAC.
  
Shouldn't Icecast scale both vCPUs? Is that the expected behaviour?   
  
 Best,
 Thiago


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180522/23b1b995/attachment.html>

Hello,
  
 At a test VPS running Ubuntu 16.04 LTS I did this:
  	sudo apt-get update 	sudo add-apt-repository ppa:certbot/certbot 	sudo 
apt-get install certbot 	sudo apt-get install icecast2 	sudo certbot 
certonly --standalone -d domain.com 	cat cert.pem privkey.pem | sudo tee 
/etc/icecast2/icecast.pem 	sudo vi /etc/icecast2/icecast.xml 
	<listen-socket>
	    <port>8443</port>
	    <ssl>1</ssl>
	</listen-socket>
	<ssl-certificate>/etc/icecast2/icecast.pem</ssl-certificate> 	sudo
chown
icecast2:icecast /etc/icecast2/icecast.pem 	sudo /etc/init.d/icecast2 
restart 
  
 After that, I have this at the log:
  
 [2018-08-12  01:47:07] INFO stats/_stats_thread stats thread started
[2018-08-12  01:47:07] INFO main/main Icecast 2.4.2 server started
[2018-08-12  01:47:07] INFO connection/get_ssl_certificate No SSL 
capability
[2018-08-12  01:47:07] INFO yp/yp_update_thread YP update thread started
  
  
 I tried restart the VPS and a lot of things (change the order of the pem 
creation, etc), all with no success.
  
 What I am doing wrong?
  
  
 Best,
 Thiago

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180811/680c713f/attachment.html>

Hi,


On 08/12/2018 02:04 AM, subscription at nextdial.com.br
wrote:
> Hello,
>  
> At a test VPS running Ubuntu 16.04 LTS I did this:
>
>  1. sudo apt-get update
>  2. sudo add-apt-repository ppa:certbot/certbot
>  3. sudo apt-get install certbot
>
You'll need an additional step at this point, see below.

>  1. sudo apt-get install icecast2
>  2. sudo certbot certonly --standalone -d domain.com
>  3. cat cert.pem privkey.pem | sudo tee /etc/icecast2/icecast.pem
>  4. sudo vi /etc/icecast2/icecast.xml
>  5. <listen-socket>
>         <port>8443</port>
>         <ssl>1</ssl>
>     </listen-socket>
>    
<ssl-certificate>/etc/icecast2/icecast.pem</ssl-certificate>
>  6. sudo chown icecast2:icecast /etc/icecast2/icecast.pem
>  7. sudo /etc/init.d/icecast2 restart
>
>  
> After that, I have this at the log:
>  
> [2018-08-12  01:47:07] INFO stats/_stats_thread stats thread started
> [2018-08-12  01:47:07] INFO main/main Icecast 2.4.2 server started
> [2018-08-12  01:47:07] INFO connection/get_ssl_certificate No SSL
> capability
> [2018-08-12  01:47:07] INFO yp/yp_update_thread YP update thread started
>  
This is because Debian (and Ubuntu, as they recycle the same packaging)
refuse to compile Icecast (and other software) with openSSL support for
political reasons.

> What I am doing wrong
>
You didn't do anything wrong as such. You just didn't know that there
are additional steps if you need TLS support.
The official Xiph.org packages are built with openSSL support:
https://wiki.xiph.org/Icecast_Server/Installing_latest_version_(official_Xiph_repositories)

$ curl
https://download.opensuse.org/repositories/multimedia:/xiph/xUbuntu_16.04/Release.key
>/tmp/multimedia-obs.key
$ gpg /tmp/multimedia-obs.key

It should yield:
pub rsa2048 2017-11-21 [SC] [expires: 2020-01-30]
0E313DB7936B4E76E720065B77EC2301F23C6AA3
uid multimedia OBS Project

$ sudo apt-key add /tmp/multimedia-obs.key
$ sudo sh -c "echo deb
http://download.opensuse.org/repositories/multimedia:/xiph/xUbuntu_16.04/
./ >>/etc/apt/sources.list.d/icecast.list"
$ sudo apt-get update
$ sudo apt-get install icecast2

Make sure it downloads the package from an OBS address and not from an
ubuntu.com or mirror. (Newer versions like 18.04 require an explicit
version or other tricks at the moment: sudo apt-get install
icecast2/2.4.2-2 )

At this point your server should already be running the Xiph.org build
of Icecast and port 8443 should answer to HTTPS.

If you would prefer to listen to the standard port of 443, please follow
these additional directions:
http://lists.xiph.org/pipermail/icecast/2015-February/013198.html
Under no circumstances you should try to reverse proxy Icecast 2.4.x -
while one can make it mostly work, it is far from trivial to set things
up in a way that will avoid most of the corner cases. Most famously,
taking down your webserver.

Cheers,

Thomas

Hello,
 We are using url authentication and getting the item bellow at the log:
  
 [2018-08-24  23:26:05] INFO auth_url/url_add_listener client auth 
(http://127.0.0.1:8001/listener/add) failed with ""
  
 I put try/catch and reviewed all my code ensure always return a proper 
status and header error code, so what trigger this kind of error?

  
  
 Regards,
 Thiago
  

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/icecast/attachments/20180824/025af30d/attachment.html>