Hosting a private python package manager in Azure or AWS

[leviya]

I work in a small team of Python developers and our aim is to create a private package manager to store our packages. I came cross pypicloud and following its tutorial I've successfully been able to upload and install packages stored in AWS-S3. That's fantastic.

However, the private package manager is launched and hosted locally typing pserve server.iniin the terminal. The package manager is then accessible at http:/ /0.0.0.0:6543/#/. Ideally, I want this server to be:

1. hosted in the cloud (possibly Azure but AWS is fine)
2. up and running 24/7
3. secure. Users must authenticate in order to see the package manager and the packages.
4. not hosted in a operative system but something higher level (e.g. docker container or Azure function app?) since it would be much easier to maintain, geo-replicate etc

How to get a secure and private server (python package manager) always up and running in the cloud?

[johnson267]

• https://docs.microsoft.com/en-us/azure/app-service/containers/app-service-linux-intro
• https://docs.microsoft.com/en-us/azure/app-service/containers/quickstart-custom-docker-image

This is roughly just a simple way to deploy a container without the burden of Kubernetes. ACS might be best if you really want a dedicated container solution (there is several orchestrators).

For the authentication part, htaccess file with basic authentication is supported by pip. This allow to do something like extra-index-url = https://login:password@pypi.myserver.com/simple/. I don't have a prefered tutorial, but google/bing something like "pip private repository" you will have no issue to find answers.

(full disclosure, I work at MS in the Azure Python SDK team)

[boblewis1857]

It is not working for me.

[stevediaz]

Hello

you can follow below steps-

1. Choose a cloud provider: Select a cloud provider such as Azure or AWS, based on your preference and requirements.

2. Set up a virtual machine or container: Create a virtual machine or container instance in the cloud environment. You can use Azure Virtual Machines or AWS EC2 for this purpose.

3. Install and configure the necessary dependencies: Install Python, pypicloud, and any other dependencies required for your package manager on the virtual machine or container.

4. Secure the server: Configure firewall rules and network security groups to restrict access to the server. Ensure that only authorized users can connect to the server.

5. Set up HTTPS: Obtain an SSL certificate and configure HTTPS for secure communication with the server. This will ensure that data transferred between clients and the package manager is encrypted.

6. Implement authentication: Configure authentication mechanisms, such as username/password or API keys, to restrict access to the package manager. This can be done using frameworks like Flask or Django.

7. Containerize the application (optional): If you prefer a containerized solution, you can create a Docker image containing your package manager and deploy it to a container service like Azure Container Instances or AWS Elastic Container Service.

8. Deploy and monitor: Deploy your package manager to the cloud server and set up monitoring to ensure it is always up and running. Configure auto-scaling and load balancing if needed.

Hope it helps you. 

Thank you .