Short brief

This documentation is written for uploading AMS to S3 and concatenating same Stream ID MP4 files.

How does Stream concatenation and uploading to S3 work?

The process begins when a new recorded stream is uploaded to S3. After the upload is completed, Ant Media Server writes the Stream ID to the following file /home/ubuntu/scripts/streamListsQueue.txt using the processStreams.sh script.

Ant Media Server Configuration Processes

To configure Ant Media Server, we need to add the settings.muxerFinishScript=/home/ubuntu/scripts/processStreams.sh parameter in /usr/local/antmedia/webapps/{Application-Name}/WEB-INF/red5-web.properties.

Here are AWS S3 Configuration and User-Defined scripts

Please ensure that your script is in the correct path and that MP4 recording is enabled in the application settings.

Instance Configuration Processes

In the /etc/crontab file, we have the concatenate_mp4.sh script. The script works every 3 minutes with a cron job: */3 * * * * root /bin/bash -l -c /home/ubuntu/scripts/concatenate_mp4.sh > /home/ubuntu/scripts/myErrorLog.txt 2>&1.

This means concatenate_mp4.sh script works every 3 min with a cron job.

You need to add the following codes for the S3 Mount:

sudo s3fs s3Bucket /mnt/s3/ -o allow_other -o default_acl='public-read'

Let me explain how the concatenate_mp4.sh script process works:

1. Check if any process of the concatenate_mp4.sh script is already running. If there is a running process, the process will finish.
2. Check the first line in streamListsQueue.txt. If it's empty, the process will exit.
3. Get the Stream ID in the first item of streamListsQueue.txt.
4. Check if the original file belongs to the Stream ID exists. If streamID.mp4 does not exist, the process will finish.
5. The script checks if streamId_1.mp4, streamId_2.mp4, streamId_3.mp4, etc. files exist. If a stream file exists, the concatenate process begins in the concatenate_files function. The concatenate process also works on the S3 mount.
6. The script tries 3 more files if they exist. If the files do not exist, the process will finish and remove the Stream ID in streamListsQueue.txt.

Finally, if the concatenate process works normally, the script will write the logs in concatenated_streams.txt. For example: /mnt/s3_n/vod/streams/677658101701365982735382_1.mp4 concatenated original file -> /mnt/s3_n/vod/streams/677658101701365982735382.mp4

Here is an example usage video:

Note: Please use the S3 Fuse folder for concatenating.