Search results

The server described in the article was tested to use 1000 subscribers simultaneously. More powerful server (96 cores for example) may pass the test for 2000 subscribes, but it should be tested. If you need more subscribers, you should deploy a CDN. For 10000 subscribers, you need at lease 5...

Hello Please send SSH sudo access to affected Edge server using this form and point to date and time of incident Our engineers will check. If you have any troubles with SSH access, please gather a report and send the report using the same form...

Good day. This depends on server hardware configuration. Actually, a load tests should be performed to check how many subscribers or publishers can connect simultaneously with a proper video quality: What kind of server do you need to run a thousand WebRTC streams?

Если статус INTERNAL_SIP_ERROR стабильно воспроизводится, воспроизведите проблему, соберите полный отчет: Сбор отладочных логов при помощи скрипта report.sh и пришлите, используя эту форму.

If you have two PBX domains, and you want to connect one of them to make a call, you should use RegistrationStatusEvent REST hook to check if connection to PBX is successful. Your backend receives ConnectionStatusEvent then RegistrationStatusEvent if PBX is available and ConnectionStatusEvent...

We see a lot of Failed to allocate port messages in server logs after WCS startup. Seems like on the server instance startup the network stack is started later than WCS: In this case WCS cannot bind media ports because the network stack is not fully active yet. You should restart WCS manually...

Good day. You can achieve this as follows: 1. Publish a stream which should be played by viewers 2. Capture VOD stream from a file: Capturing VOD from a file 3. Inject VOD stream to the playback stream: Injection implementation in build 5.2.1618 and newer Please note that stream to inject to and...

Check if some nginx configs contains listen 80 directive grep -r listen /etc/nginx/* Please read also this post.

Yes, you can play the same published stream using WebRTC and LL HLS simultaneously. For example, you can play WebRTC on desktop PC or Mac and LL HLS on mobile device. A frontend should decide what technology to use.

Please check if the port is busy by other application: sudo netstat -nlp Please also check nginx configuration: there should not be any server entities in nginx.conf listening for http port 80 unless they really needed.

Такое может быть при большой нагрузке на сервер: встроенный бэкенд не успевает обрабатывать REST хуки, либо превышено количество одновременных соединений. В этом случае предусмотрена повторная отправка запроса, по умолчанию повторяется три раза rest_client_request_retry_count=3 Кроме того, в...

We recommend you to test an encrypted SSL private key on staging server before moving it to production.

This should work for encrypted SSL private key too. You should pass a passphrase to nginx as recommended in this post.

nginx displays the following error in /var/log/nginx/error.log connect() to 192.168.0.223:8080 failed (13: Permission denied) while connecting to upstream This means SELinux disables a local http connections. The following command enables them: setsebool httpd_can_network_connect on -P After...

Please provide more screenshots with error output and ~/.ssh/config file you modified.

Hello It seems Low Latency HLS would be better. You would need a HLS player like HLS.js or Video.js See more: https://docs.flashphoner.com/display/WCS52EN/In+a+browser+via+HLS#InabrowserviaHLS-LowLatencyHLSsupport As you can see, HLS player does not require Flashphoner Web SDK at all.

We checked the report. You've set a wrong option So it does not work properly. Please set rest_client_request_retry_count=10 Also you've not collected a system configuration, so we can't check it. Please collect a proper report sudo /usr/local/FlashphonerWebCallServer/tools/report.sh --sysinfo...

Now it seems like port 443 is opened at server side but blocked by router Please check your NAT configuration.

We checked the report. Seems like you've removed some files (jstack.log, Java heap dump) from the report. Also, you've replaced some info. Please, don't do this again. The problem is lack of Java heap memory: You're using a default Java heap size (1 Gb) This may be not enough under a high...

Seems like the whole server is down, so we cannot check