This project provides installation scripts and configuration files for deploying CSGHub instances, including Helm Chart and Docker Compose scripts, to simplify the deployment process in various environments.
Please go to here for more details information about CSGHub.
Choose the right method for your deployment:
【Recommend for quick-start】 If you are trying CSGHub for fast-try on your test/dev environment without complicated configs, you could go to the Docker section
【Recommend for production-env】 If you are using CSGHub for production envirement with kubernetes supported, you should go to the HelmChart section
If you are familar with Docker Compose and trying to use CSGHub, you could go to the DockerCompose section
- The docker deployment method is mainly used for simple functional experience testing. It has just been launched and some functions are not yet perfect. They will be supplemented later.
- Functions that depend on k8s are not yet completed.
- For mor details about docker installation, please refer to docker
- The helm chart method is suitable for scenarios with high stability and availability, such as production environments.
- helm chart only supports
gitaly
as the git server backend,gitea
is not supported. - For more details about helm chart installation and deployment, please refer to helm-chart
- compose mode can be used for test and develop purpose. It is recommended to use helm chart installation for production environments.
- CSGHub instance that deployed with compose mode cannot directly use functions which rely on the kubernetes platform, such as space, model inference, and model fine-tuning. Kubernetes's deployment and configuration are not within the scope of the compose installation method, it needs further manual configurations which can be found here
- Starting from CSGHub v0.9.0, CSGHub no longer provides continuous support for gitea backend, and it is recommended to use
gitaly
as default git server backend. - Provide a solution for one-click deployment to Alibaba Cloud, Deployment Link
- For more details about compose installation and deployment, please refer to docker-compose