Currently using v2.2.4.

A minor disruption in our live feed to our server with Centvoa, continues to lose about half our listeners on average.

We have one server, and one feed, there are no other clients set up with Centova.

So we have accessed /centovacast/system/config.php and defined a 10 second delay here:

define('SERVSTATE_CACHE_TIMEOUT',10);

But this setting (10 seconds) seems to have no effect.

So each time we even every-so-briefly lose connection, when it is back up the Nicecast app just comes back with a series of "login failures" for the next 5 to 10 minutes, sometimes seemingly longer. Then finally we are allowed to reconnect, but by then we lose 50% or more of our listeners.

So what does it take exactly for Centova to "look" for our live feed, and log us back in right away, or with a minimum delay of less than one minute when the connection is disrupted?

Any help here would be vastly appreciated, we have been struggling with this same issue for close to 2 years now.

Hello Joseph,

Unfortunatelly this has nothing to do with Centova Cast, once Shoutcast/Icecast have been configured and launched they are pretty much on their own.

By messing up with Centova Cast's cronjob or config files you will only affect the behavior of the control panel and the autodj source but not the server itself.

While dealing with network failure, the best way is to set up a fallback message, using Icecast you can make use of the fallback mountpoint while in shoutcast you'll need to upload an ondemand audio file.

Note: The fallback mountpoint in icecast is your best bet, you can get creative and connect to multiple mountpoints using multiple ISPs to maximize availability if it's really that important.

You may also want to turn on the "kick idle source option" (not sure if Icecast supports this), and the reconnect feature on your source software.

Hope this helps.

Thanks for your reply.

Okay, so what Nicecast logs into (to upload the source/remote stream for broadcast) is then apparently the Shoutcast server? And then Centovacast get's the stream from there?

Okay, if this is true, why do you suppose that Shoutcast does not let us log back in for 10 minutes, etc. once our connection to the internet is briefly disrupted for just a few seconds? And how can we speed up the interval that Shoutcast uses to let us log back in?

Question:

In the following file:
/home/centovacast/vhosts/audioart/etc/server.conf

... I have found this:

AutoDumpSourceTime=300

I think this was my problem, yes? Okay, so now I have lowered this value to 15. What do I need to do in order to put this into effect? Do I need to "Restart Server"?

Question 2:

In this file:
/home/centovacast/shoutcast/sc_serv.conf

I have found this:

AutoDumpSourceTime=30

I have also changed this value to 15

What do I need to do to also put THIS into effect?

And am I on the right track here?

Thank you.

This has already being answered in ticket HT8-OWP9-873, and in my first reply to this topic (except for the restart part).

About the restart, yes you should always restart after any changes for the configuration to take effect.

Also, you've mentioned some kind of load balance, this is probably the worst thing you can do when streaming live, since it will change your IP address and shoutcast will not recognize your current session.

If you'd like to use multiple ISPs for redundancy try using Icecast's fallback mountpoint feature as mentioned above.