I’m deploying Openvidu Pro (2.20.0) on premise with docker-compose.
The services are up and running well, the individual recordings are started correctly.
However, when I stop the recording ( https:// YOUR_OPENVIDUSERVER_IP /openvidu/api/recordings/stop/ RECORDING_ID request), Openvidu server deletes webm file and includes only json file into recording zip. Please, advice how to preserve the recorded video
Deletion of recorded files in the Media Nodes is the expected behavior. Your Master Node will download video files from Media Nodes after recordings are stopped, so all of the files are centralized in the Master Node. Once properly donwloaded, it will delete them from Media Nodes so no disk space is wasted with duplicated files.
I can barely see your logs, but I think they state:
File [...]str_CAM_BVv4_con_S1bqqMAvxF.webm is 7184 KB in size
File successfully downloaded from http:/173.201.19.150:3000/.../str_CAM_BVv4_con_S1bqqMAvxF.webm. Deleting it
File located at http:/173.201.19.150:3000/.../str_CAM_BVv4_con_S1bqqMAvxF.webm successfully deleted
This is the expected normal behavior. Now, if in your .zip file located at your Master Node recordings folder you do not see the video file str_CAM_BVv4_con_S1bqqMAvxF.webm, then another problem must be taking place. In this case, we need further information. Probably the best source of intel here is more logs of openvidu-server. Please, provide to us the complete log of this recording, starting before the lines you posted above, and ending after them. From the beginning of the recording to well after the stopping process has completed.
Thank you for the quick reply. Yes, the zip in media-node folder is also empty. Here is the full log:
The log you provided is from a different recording, isn’t it?
Do you get this same behavior in all of the INDIVIDUAL recordings you start? Are you always able to reproduce the lack of webm files in the ZIP?
Yes, the behavior is the same for all individual recordings. Webm files lack both in cases when the recording is stopped manually or by timeout.
I can’t find anything wrong in your log. It is the same as the log of a healthy recording.
I am starting to think that the problem must be related to something specific to your particular infrastructure (tool or path permissions?, disk space?)
Please, provide as much information as possible about your OpenVidu Pro on premises deployment, so we can have a more complete view of your environment.
If you want you can contact us privately via pro.support@openvidu.io to provide us temporal access to your deployment so we can test it ourselves.