Support edge/forward to Aliyun and Tencent CDN.
winlinvip opened this issue · comments
Fixed in 2.0.255 and 3.0.40
First, create live domain at here, please select your region.
First, click here to create a domain for live streaming, and select your region.
Click manage button to mange the domain, about DRM and url.
Click the "Manage" button to manage the domain, configure anti-theft chain, streaming, and playback information.
TRANS_BY_GPT3
Then, config the DRM and publish/play url of that domain.
Then, configure the DRM (authentication information) of the domain name and generate URLs for streaming and playback.
For each publish/play url, you should generate the auth_key. It's better to auto generate url by algorithm.
For each streaming and playback URL, the authentication information needs to be generated separately. The best practice is to generate URLs programmatically, and for the specific algorithm, please refer to the live streaming authentication documentation.
TRANS_BY_GPT3
Next, start SRS in edge or forward server:
# Forward to Aliyun CDN.
listen 1935;
max_connections 1000;
daemon off;
srs_log_tank console;
vhost __defaultVhost__ {
forward xxxx.alivecdn.com;
}
# Edge of Aliyun CDN.
listen 1935;
max_connections 1000;
daemon off;
srs_log_tank console;
vhost __defaultVhost__ {
mode remote;
origin xxxx.alivecdn.com;
}
Please replace the xxxx.alivecdn.com as your publish domain in url:
Replace xxxx.alivecdn.com with your streaming domain name, as shown in the following figure:
TRANS_BY_GPT3
Finally, publish stream to Aliyun CDN, or your SRS then play:
# Publish to Aliyun CDN
ffmpeg -re -i doc/source.200kbps.768x320.flv -c copy -y -f flv -y \
'rtmp://xxxx.alivecdn.com/live/livestream?vhost=xxxx&auth_key=xxxx'
# Publish to your SRS
ffmpeg -re -i doc/source.200kbps.768x320.flv -c copy -y -f flv -y \
'rtmp://127.0.0.1/live/livestream?vhost=xxxx&auth_key=xxxx'
Play stream rtmp://xxxx/live/livestream?auth_key=xxxx: http://www.ossrs.net/players/srs_player.html?app=live&stream=livestream&server=xxxx&port=1935&autostart=true&vhost=xxx
Today, forwarding with srsv2.0-r5 to Aliyun was not successful. There might be some configuration errors.
vhost __defaultVhost__ {
forward video-center.alivecdn.com;
}
Streaming center address:
rtmp://video-center.alivecdn.com/AppName/StreamName?vhost=xxxx //accelerated domain
If using OBS or ffmpeg -re -i to directly stream to this streaming center address, it can be played.
ffmpeg -re -i /home/redhat/xlopening.mp4 -c copy -y -f flv -y 'rtmp://127.0.0.1/live1007/live1007st?vhost=xxxxx'
ffmpeg -re -i /home/redhat/xlopening.mp4 -c copy -y -f flv -y 'rtmp://video-center.alivecdn.com/live1007/live1007st?vhost=xxxxx'
However, if streaming to SRS and then forwarding to Alibaba Cloud, playback will not be possible. Does this mean that Alibaba Cloud no longer supports this operation? The stream is also shown as offline in the Alibaba Cloud backend. The console displays the following error message: 1007 reset by peer.
[2018-11-16 09:15:11.795][trace][19807][100] RTMP listen at tcp://0.0.0.0:1935, fd=7
[2018-11-16 09:15:11.795][trace][19807][100] signal installed
[2018-11-16 09:15:11.806][trace][19807][100] http: api mount /console to ./objs/nginx/html/console
[2018-11-16 09:15:11.816][trace][19807][100] ingest thread cid=103, current_cid=100
[2018-11-16 09:15:12.169][trace][19807][105] RTMP client ip=x.1xx.x.102
[2018-11-16 09:15:12.169][trace][19807][105] srand initialized the random.
[2018-11-16 09:15:12.217][trace][19807][105] simple handshake success.
[2018-11-16 09:15:12.217][trace][19807][105] connect app, tcUrl=rtmp://xx.xx.xx.xx/live, pageUrl=, swfUrl=, schema=rtmp, vhost=xx.xx.xx.xx, port=1935, app=live, args=null
[2018-11-16 09:15:14.244][trace][19807][105] client identified, type=fmle-publish, stream_name=0, duration=-1.00, param=
[2018-11-16 09:15:14.244][trace][19807][105] connected stream, tcUrl=rtmp://xx.xx.xx.xx/live, pageUrl=, swfUrl=, schema=rtmp, vhost=__defaultVhost__, port=1935, app=live, stream=0, param=, args=null
[2018-11-16 09:15:14.265][trace][19807][105] source url=/live/0, ip=1.198.22.102, cache=1, is_edge=0, source_id=-1[-1]
[2018-11-16 09:15:14.340][trace][19807][105] start forward rtmp://xx.xx.xx.xx:1935?vhost=__defaultVhost__ to rtmp://video-center.alivecdn.com:1935?vhost=__defaultVhost__, tcUrl=rtmp://video-center.alivecdn.com/live, stream=0
[2018-11-16 09:15:14.350][trace][19807][105] forward thread cid=108, current_cid=105
[2018-11-16 09:15:14.360][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
[2018-11-16 09:15:14.360][trace][19807][105] start publish mr=0/350, p1stpt=20000, pnt=5000, tcp_nodelay=0, rtcid=109
[2018-11-16 09:15:14.376][trace][19807][108] complex handshake success.
[2018-11-16 09:15:14.439][trace][19807][108] connected, version=, ip=, pid=0, id=0, dsu=1
[2018-11-16 09:15:14.463][trace][19807][105] got metadata, width=1280, height=720, vcodec=2
[2018-11-16 09:15:14.463][trace][19807][105] protocol in.buffer=0, in.ack=0, out.ack=2500000, in.chunk=1440, out.chunk=60000
[2018-11-16 09:15:14.463][trace][19807][105] 34B video sh, codec(7, profile=Main, level=3.1, 1280x720, 0kbps, 0fps, 0s)
[2018-11-16 09:15:14.501][trace][19807][108] protocol in.buffer=0, in.ack=5000000, out.ack=2500000, in.chunk=4096, out.chunk=60000
**[2018-11-16 09:15:14.510][error][19807][108][104] recv server control message failed. ret=1007(Connection reset by peer)**
[2018-11-16 09:15:17.519][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
[2018-11-16 09:15:17.535][trace][19807][108] complex handshake success.
[2018-11-16 09:15:17.597][trace][19807][108] connected, version=, ip=, pid=0, id=0, dsu=1
[2018-11-16 09:15:17.605][trace][19807][108] protocol in.buffer=0, in.ack=5000000, out.ack=2500000, in.chunk=4096, out.chunk=60000
**[2018-11-16 09:15:17.943][error][19807][108][104] recv server control message failed. ret=1007(Connection reset by peer)**
[2018-11-16 09:15:20.952][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
[2018-11-16 09:15:20.968][trace][19807][108] complex handshake success.
[2018-11-16 09:15:21.032][trace][19807][108] connected, version=, ip=, pid=0, id=0, dsu=1
[2018-11-16 09:15:21.040][trace][19807][108] protocol in.buffer=0, in.ack=5000000, out.ack=2500000, in.chunk=4096, out.chunk=60000
**[2018-11-16 09:15:21.374][error][19807][108][104] recv server control message failed. ret=1007(Connection reset by peer)**
[2018-11-16 09:15:24.381][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
[2018-11-16 09:15:24.392][trace][19807][108] complex handshake success.
[2018-11-16 09:15:24.447][trace][19807][108] connected, version=, ip=, pid=0, id=0, dsu=1
[2018-11-16 09:15:24.453][trace][19807][108] protocol in.buffer=0, in.ack=5000000, out.ack=2500000, in.chunk=4096, out.chunk=60000
**[2018-11-16 09:15:24.779][error][19807][108][104] recv server control message failed. ret=1007(Connection reset by peer)**
[2018-11-16 09:15:27.784][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
[2018-11-16 09:15:27.794][trace][19807][108] complex handshake success.
[2018-11-16 09:15:27.847][trace][19807][108] connected, version=, ip=, pid=0, id=0, dsu=1
[2018-11-16 09:15:27.852][trace][19807][108] protocol in.buffer=0, in.ack=5000000, out.ack=2500000, in.chunk=4096, out.chunk=60000
**[2018-11-16 09:15:28.154][error][19807][108][104] recv server control message failed. ret=1007(Connection reset by peer)**
[2018-11-16 09:15:31.162][trace][19807][108] forward connected, stream=0, tcUrl=rtmp://xx.xx.xx.xx/live to server=video-center.alivecdn.com, port=1935
TRANS_BY_GPT3
Actually, I just found out that I don't need to stream, just starting SRS results in an error.
recv server control message failed. ret=1007(Connection reset by peer)
TRANS_BY_GPT3
I tried it, it is possible. The SRS configuration is as follows:
listen 1935;
max_connections 1000;
daemon off;
srs_log_tank console;
vhost __defaultVhost__ {
forward video-center-bj.alivecdn.com;
}
Generate the streaming address in the Alibaba Cloud console:
When streaming to SRS, it will forward to Alibaba Cloud. Please note that when streaming, it is necessary to include authentication information in the URL.
ffmpeg -re -i doc/source.200kbps.768x320.flv -c copy -f flv \
-y 'rtmp://127.0.0.1/live/livestream?vhost=live.aliyun.ossrs.net&auth_key=xxx'
When playing, it is also necessary to include authentication information in the address.
TRANS_BY_GPT3
May I ask a question? Currently, I am pushing the stream to SRS, then using the configuration file to forward it to an Alibaba Cloud server. However, the online stream is not displayed on the Alibaba Cloud server, even though the resource monitoring in the domain management section shows resource movement. The videos cannot be played. Can you please explain why?
TRANS_BY_GPT3
Difference between distributing to backup SRS and distributing to Alibaba Cloud
TRANS_BY_GPT3
[2020-06-08 20:34:39.382][Trace][9011][217] connected, version=0.0.0.0, ip=, pid=0, id=0, dsu=1
From here, things are different.
TRANS_BY_GPT3
ffmpeg -rtsp_transport tcp -re -i rtsp://admin:luke1234@192.168.0.108 -vcodec copy -acodec copy -f flv rtmp://192.168.0.22/AppName/StreamName?auth_key=1591621430-0-0-ca4f580a543d387b7874ea7baff0db8f
Streaming command
TRANS_BY_GPT3
ffmpeg -rtsp_transport tcp -re -i rtsp://admin:luke1234@192.168.0.108 -vcodec copy -acodec copy -f flv rtmp://192.168.0.22/AppName/StreamName?auth_key=1591621430-0-0-ca4f580a543d387b7874ea7baff0db8f
Streaming command
Inside the log, there is information that SRS actually forwarded successfully.
[2020-07-30 10:26:56.308][Trace][10329][744] <- EPB time=10000366, msgs=25, okbps=159,0,0, ikbps=2,0,0
[2020-07-30 10:27:06.320][Trace][10329][744] <- EPB time=20000426, msgs=26, okbps=309,0,0, ikbps=1,0,0
Assuming that the push/pull streaming domain names for Alibaba are: push.ali/pull.ali
To push the stream, you simply need to push it as follows: rtmp://push.ali/AppName/StreamName?auth....
When pulling the stream, you need to add vhost=push domain name, and this way you can successfully pull the stream.
rtmp://pull.ali/AppName/StreamName?vhost=push.ali&auth=......
The above information is based on SRS 3.0.141.
TRANS_BY_GPT3
May I ask, I am currently streaming to SRS and then forwarding it to an Alibaba Cloud server through a configuration file. The online stream is not displayed on the Alibaba Cloud server, but the resource monitor in the domain management shows that the resources are flowing. The video cannot be played. Can you please tell me why?
I am currently facing the same situation as you. Could you please let me know how you resolved this issue at that time?
TRANS_BY_GPT3
@liguocahng
Push normally,
When pulling, you need to add the vhost parameter to the pulling address, such as: rtmp://pull.ali/AppName/StreamName?vhost=push.ali&auth=......
After adding the vhost parameter, it only affects the ability to pull and watch the stream, but in reality, the vhost parameter may also impact other functions of Alibaba Cloud, such as missing statistical information or inability to generate recording files.
TRANS_BY_GPT3
【Stream to SRS and forward to Alibaba Cloud CDN Live Center】
Today, while testing version 3.0.156, we found that:
Streaming to SRS using a domain and forwarding to CDN failed.
Streaming to SRS using an IP and forwarding to CDN was successful.
The error message for the domain failure is: [Please provide the error message for the domain failure.]
TRANS_BY_GPT3
[Stream to SRS -> Forward to Alibaba Cloud CDN Live Center]
Today, while testing version 3.0.156, we found:
Pushing stream to SRS using a domain name and forwarding to CDN failed.
Pushing stream to SRS using an IP address and forwarding to CDN was successful.Error message for domain name failure:
Open a new issue to describe the problem of domain name forwarding, and provide a detailed description.
TRANS_BY_GPT3
In some other cases, when forwarding, the vhost is also carried and cannot be removed through configuration, causing similar issues.
Expectation: The configuration of forward should allow the cancellation of vhost parameter configuration, specifically only vhost, while other URL query parameters should still be used.
Temporarily using exec to resolve it.
exec {
enabled on;
publish ./objs/ffmpeg/bin/ffmpeg -f flv -i rtmp://127.0.0.1:[port]/[app]?vhost=[vhost]/[stream] -c copy -f flv rtmp://cdn.con/[app]/[stream];
}TRANS_BY_GPT3
@springjk
Streaming to SRS on my end and forwarding to the Alibaba Cloud CDN Live Center aims to increase the concurrency of live broadcasting. How does adding ffmpeg forwarding affect machine performance, stability, and other factors?
Another solution is to modify the source code.
TRANS_BY_GPT3
@langyastudio
Using this virtual machine as an example:
[root@localhost ~]# cat /proc/cpuinfo | grep name | cut -f2 -d: | uniq -c
16 Intel(R) Xeon(R) CPU E7-8850 v2 @ 2.30GHz
Currently, there are five live streams, transcoded to 720p and 480p, then pushed to CDN using ffmpeg. The load is between four and five.
In my scenario, the resource usage is roughly similar to when I used "forward" before. The resource usage also depends on the incoming bitrate and the number of transcoding processes, but I haven't done any specific calculations. As for stability, I haven't noticed any issues in the past few days. I see that "forward" also uses ffmpeg for forwarding, so there shouldn't be much difference.
TRANS_BY_GPT3