Wowza Community

multicast stream and "Stream not healthy [stream startup timeout]"

Hi,

I have a Wowza 4.0.3 installed on a 6.5 Centos with 2 bonded network cards.

bond0 has a private IP address 172.16.1.51 and get multicast stream from a transcoder DVEO Brutus IP Telco.

The transcoder send multicast RTP with transport stream envelope.

bond1 has a public IP address for clients retransmission of channels.

Multicast stream is coming from transcoder to Wowza server on bond0, here is the tcpdump output :

17:32:36.558548 IP 10.30.2.39.3001 > 239.200.2.39.3001: UDP, length 1328

and netstat -g output :

bond0           1      239.200.2.39

I cutomized VHost.xml as follow :

RTP/DatagramConfiguration/Incoming :

<MulticastBindToAddress>true</MulticastBindToAddress>
<MulticastInterfaceAddress>172.16.1.51</MulticastInterfaceAddress>
<TrafficClass>0</TrafficClass>

And Application.xml :

<Properties>
<Name>multicastInterfaceAddress</Name>
<Value>172.16.1.51</Value>
<Properties>

I created a stream (multicast.stream) on a Live application (LiveFromTranscoder) with IP : udp://239.200.2.39:3001 and RTP MediaCaster Type.

But streaming to flash player isn’t working, and I’ve got “Stream not healthy [stream startup timeout]” errors from logs :

2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	RTPMediaCaster.Reconnector[1072576560:LiveFromTranscoder/_definst_:multicast.stream]: start: 232	-	-	4825.496	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	RTPMulticastListener.run(stop)	_defaultVHost_	LiveFromTranscoder	_definst_	4825.499	-	80null	127.0.0.1	rtsp	-	known	599970080	0	0	-	-	-	-	-	-	-	-	-	-	-	-	-	null	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	RTPSessionDescriptionDataProviderBasic.getStreamInfo[LiveFromTranscoder/_definst_]: URI: udp://239.200.2.39:3001	4825.597	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:38:42	CEST	create	stream	INFO	200	-	-	_defaultVHost_	LiveFromTranscoder	_definst_	0.0	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	232	0	0	0	multicast.stream	-	-	-	-	-	null	null	-	null	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	RTPDePacketizerMPEGTS.init: Drop incomplete video frames	_defaultVHost_	LiveFromTranscoder	_definst_	4825.6	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	-	-	-	-	-	-	-	-	-	-	-	-	null	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	MulticastTransport.bind[LiveFromTranscoder/_definst_]: 239.200.2.39/3001	_defaultVHost_	LiveFromTranscoder_definst_	4825.601	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	-	-	-	-	-	-	-	-	-	null	-
2014-08-01	17:38:42	CEST	publish	stream	INFO	200	multicast.stream	-	_defaultVHost_	LiveFromTranscoder	_definst_	0.003	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	232	0	0	0	multicast.stream	-	-	-	-	-	null	null	-	null	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	RTPMediaCaster.Reconnector[1072576560:LiveFromTranscoder/_definst_:multicast.stream]: done: 232	-	-	-	4825.602	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:38:42	CEST	comment	server	INFO	200	-	MediaCasterStreamValidator.getMonitor[LiveFromTranscoder/_definst_] create:multicast.stream	-	-	-	4825.996	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	MediaCasterStreamValidator.onValidateMediaCaster[LiveFromTranscoder/_definst_] Stream not healthy [stream startup timeout]: mp4:multicast.stream	-	-	-	4846.004	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	MediaCasterStreamValidator.onValidateMediaCastersStop[LiveFromTranscoder/_definst_] Reset stream [mediaCaster]: mp4:multicast.stream	-	-	-	4846.004	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	RTPMediaCaster.resetConnection[1072576560:LiveFromTranscoder/_definst_:multicast.stream]: 	-	-	-	4846.005	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	RTPMediaCaster.closeRTPSession[1072576560:LiveFromTranscoder/_definst_:multicast.stream]	-	-	-	4846.005	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	MulticastTransport.unbind[LiveFromTranscoder/_definst_]: 239.200.2.39/3001 sent:0 recv:0	-	-	-	4846.008	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	unpublish	stream	INFO	200	multicast.stream	-	_defaultVHost_	LiveFromTranscoder	_definst_	20.41	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	232	0	0	0	multicast.stream	-	-	-	-	-	null	null	-	null	-
2014-08-01	17:39:02	CEST	destroy	stream	INFO	200	multicast.stream	-	_defaultVHost_	LiveFromTranscoder	_definst_	20.412	-	80	null	127.0.0.1	rtsp	-	known	504519732	0	0	232	0	0	0	multicast.stream	-	-	-	-	-	null	null	-	null	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	RTPMediaCaster.Reconnector[1072576560:LiveFromTranscoder/_definst_:multicast.stream]: start: 233	-	-	4846.01	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
2014-08-01	17:39:02	CEST	comment	server	INFO	200	-	RTPMulticastListener.run(stop)	_defaultVHost_	LiveFromTranscoder	_definst_	4846.012	-	80null	127.0.0.1	rtsp	-	known	504519732	0	0	-	-	-	-	-	-	-	-	-	-	-	-	-	null	-
2014-08-01	17:39:03	CEST	comment	server	INFO	200	-	RTPSessionDescriptionDataProviderBasic.getStreamInfo[LiveFromTranscoder/_definst_]: URI: udp://239.200.2.39:3001	4846.111	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-

Do you have an idea whether I’m doing something wrong or if there is a bug ?

Thanks !!!

Romain

Romain,

‘Stream not healthy’ suggests Wowza is unable to connect to your multicast source. Have you been able to connect using VLC?

I would suggest reverting your settings to default, test to validate working multicast playback, and then proceed with testing different settings.

-Tim

Hello together.

I’m bringing this topic up again as I have the same issue.

I have also a Wowza 4.0.3 running on a CentOS 6.5 with two interfaces :). One internal: 192.168.3.150 and one for external (where the public www clients connect to). We have a Harmonic Divicom Electra 8222 which provides 2 UDP Multicasts via the internal network (192.x). In the meantime I updated to 4.1.0 as we have a perpetual license and I disabled one Multicast for debugging.

I see on the wowza server with tcpdump that the UDP packets come in from the Playout with IP 192.168.3.199 (harmonic encoder IP) as source over the multicast group 225.123.123.123:10000.

I attached an laptop to the internal network and checked with VLC that the streams are correct. And they are correct.

I also edited the two mentioned xml-files (I also tried without).

So, it’s really nearly the same setup :).

And I got also the “Stream not healthy” error.

In the stream file I tested with udp://0.0.0.0:10000, udp://127.0.0.1:10000 and also udp://225.123.123.123:10000. Nothing works.

Just one time I saw an picture!! I installed VLC on the server and used the sample.mp4 to provide a UDP Multicast from the server itself (via command-line). I used udp://239.255.255.1:10000. I added this URL to the stream file. And hurray…I saw an picture. The thing was that I recognized the Multicast was received via the interface with the public IP…not the internal one. So that was the difference.

So I had the idea to disable the interface with the external IP and try it again…with the stream from the harmonic encoder and also with the VLC on the wowza server…It didn’t work. Again a “stream not healthy”. So I had no further ideas. It was not reproducable.

As this is very strange behaviour I would kindly ask for support.

I spended a lot of time doing googling and forum search.

How is the correct configuration which such an setup described from romain.a and from me?

Receiving a UDP Multicast (of group 225.123.123.123:10000) via an internal network on interface A and provide an OTT stream via the external IP on interface B.

Thank you very much.

I hope we can resolve this issue as it has an big impact on our setup and I think it is good for the wowza documentation. It’s the use case why we bought this license some time ago ( we just had no time in the past to try this use case :).

Greetings from Berlin.

I seem to be having the same issue. I am running Wowza 4.2 on a centOS 7 box. Unicast runs fine but multicast drops out every few minutes. At first I thought it was a network issue but VLC and some packet capturing have ruled that out. I am running 2 NICs like the guys above and that dosen’t seem to be an issue. Through some testing and talking with support I was sent to this artical https://www.wowza.com/docs/how-to-reduce-interference-between-incoming-multicast-streams but to no effect.

This ran fine on my original fedora box without having to change anything from the install.

I have a support ticket open but need to get this running asap as we need the system up and running last week. If anyone found a solution to this issue please let me know. I may try a totally fresh install of the box (OS and all) if I can’t get this working. I am at the end of my rope from a troubleshooting standpoint.

Thanks

UPDATE:

The issue for me was that we had firewalld running. firewalld blocks igmp traffic by default and will keep the multicast renew requests from getting through. You won’t see this issue if VLC as VLC spams group join instead of responding to igmp requests. For those of you who are running firewalld you must add this line to let the correct traffic through:

firewall-cmd --permanent --direct --add-rule ipv4 filter INPUT 0 -p igmp -j ACCEPT

I hope this helps some of you.

I solved this issue with:

  • -Enable promiscuous mode: add net.ipv4.conf.all.rp_filter=0 to sysctl.conf
  • -Enable traffic in firewalld: firewall-cmd --permanent --direct --add-rule ipv4 filter INPUT 0 -m pkttype --pkt-type multicast -j ACCEP

Regards.