Kubernetes/Docker support

Hello,

Our design team is taking into consideration to use DaVinci Resolve for their work. I am trying to set up a POC as fallow:

- Environment Kubernetes
- Rendering remote on specific pods with horizontal scaling. The scaling will be done most probably on events.

Questions:
- Does DaVinci Resolve supports docker/Kubernetes?
- Did anyone has accomplished this task by now?
- Dose DaVinci officially supports docker/Kubernetes on the rendering side?

Thank you!

Answers: No, no, no.

It's also sort of pointless as one Resolve instance wants all the GPUs a host has to offer. Dedicated.

You may get it working in K8s, but what do you gain? GPU VM Scale Set in Azure may be way easier.

Answers: No, no, no.

It's also sort of pointless as one Resolve instance wants all the GPUs a host has to offer. Dedicated.

You may get it working in K8s, but what do you gain? GPU VM Scale Set in Azure may be way easier.

Thank you Peter. Regarding the GPU I will expose the CUDA device to docker. I will give it a try today to put the studio on a docker file and start from there.

I have set it up in headless mode in Docker, but it was a while ago. I can't remember how far I got. What are you trying to achieve by running a GUI bound app in a container?

I have set it up in headless mode in Docker, but it was a while ago. I can't remember how far I got. What are you trying to achieve by running a GUI bound app in a container?

Hi Igor.

I am trying to create a "rendering farm". The horizontal scaling will be cover by "keda" based on the RabbitMQ event but the first step is to run Davinci Resolver on docker without "desktop requirements". There is a way to run it in docker but is mounting x11 from the local machine and this is not what I want. I want Davinci to run directly on docker and I will make sure I will expose the Cuda graphic card to the docker.

Right, that kind of rings a bell. The CUDA is the tricky part. I'd love to hear an update on the project later.

Right, that kind of rings a bell. The CUDA is the tricky part. I'd love to hear an update on the project later.

We have decided not to be the "pioneer" and looking for other solutions that is supporting docker/kubernetes