Blender Render Container for Kubernetes

This is a Blender headless/remote renderer image, based on RHEL UBI, designed to run in Kubernetes environments. It pulls an asset-embedded .blend file from a webserver and uploads the output to an S3 bucket.

Prerequisites

You will need an asset-embedded .blend file on an accessible webserver. Authentication is currently NOT SUPPORTED.

You will need a Minio S3 endpoint available for the output to be stored. Other S3 endpoints currently are NOT SUPPORTED.

Example Local Usage

The following example, for testing purposes, assumes you already have an image built and hosted in an accessible registry. Specific environment variables are required:

• DEBUG: when set to 1 will use -x mode for the internal bash script
• BLEND_LOCATION: The URL of the .blend file that will be fetched for processing
• RENDER_TYPE: May ONLY be set to either image or animation
• S3_ENDPOINT: The url of the S3 API endpoint
• S3_KEY: The user/key used to authenticate to the S3 API endpoint
• S3_SECRET: The secret used to authenticate to the S3 API endpoint

podman run --rm \
-e BLEND_LOCATION=http://192.168.1.235:8080/fedora.blend \
-e RENDER_TYPE=image \
-e S3_ENDPOINT=http://192.168.1.235:9000 \
-e S3_KEY=testuser_key \
-e S3_SECRET=testuser_secret \
quay.io/openshiftdemos/blender-remote

Example Minio

If you want to quickly test locally with Minio, you can run an ephemeral Minio server with the following:

podman run -p 9000:9000 -p 9001:9001 \
  quay.io/minio/minio server /data --console-address ":9001"

The default admin is minioadmin with password minioadmin and you can use those values for your S3_KEY and S3_SECRET.

Example Webserver

If you have Node.js installed, you can use the http-server Node module to serve the examples folder of this repo:

cd examples
http-server .

Example Kubernetes usage

If you have a Kubernetes cluster handy, there is an example Job and ConfigMap that can be used. It assumes the image hosted in Quay. Modify the example-animation-configmap.yaml or example-image-configmap.yaml to suit your environment, and then:

kubectl create -f k8s/example-TYPE-configmap.yaml
kubectl create -f k8s/example-job.yaml