Skip to main content

Containerization

As a dynamic and distributed system, Celte can be run in a containerized environment. This allows for easy deployment and scaling of the system. Kubernetes is our first choice to deploy Celte in a production environment. However, for development purposes, we use docker-compose to run Celte in a containerized environment.

This is an overview of how Celte is designed to run in a dockerized environment.

Docker Pulsar​

Celte uses pulsar as a communication protocol. To run pulsar in a docker container, you can use the following command:

./automation/run --pulsar

Docker Celte SDK​

To run the Celte SDK in a docker container, you can use the following command from the root of the Celte repository:

to build the container from a ARM64 machine​

docker buildx build --platform linux/amd64 -t [NAME_OF_YOUR_IMAGE] . --output type=docker

to build the container from a x86 machine​

docker build -t [NAME_OF_YOUR_IMAGE] .

to run the container​

docker run -v $(pwd):/workdir -it [NAME_OF_YOUR_IMAGE] /bin/bash

then you can run the SDK with the following command:

cd /workdir
cmake --preset default ..
make -j

Kubernetes​

Is is up to your technical constraints to deploy Celte on Kubernetes or using another orchestrator.

Right now the K8S configuration is present inside the devops folder. Inside, you will find the following files:

├── devops
│   ├── master.deployment.yaml
│   ├── master.hpa.yaml
│   ├── server-node.deployment.yaml
│   ├── server-node.hpa.yaml

This is a basic configuration to deploy Celte on Kubernetes. You can modify it to fit your needs.

This is a configuration example for the master deployment:

spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: master
  minReplicas: 1
  maxReplicas: 4
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 60
  - type: Resource
    resource:
      name: memory
      target:
        type: Utilization
        averageUtilization: 60

The important part is the spec section. You can modify the minReplicas and maxReplicas to fit your needs.

And the minReplicas and maxReplicas to control the autoscaling of the master.

The current image of the Celte master is clmt/celte-master:latest.