I’m just starting setting up my dev environment for algorand.
My goto way of installing new services is usually via docker.
The docker “install” of the docs seems to be a bit odd compare to the usual ways of spinning up containerized services.
Has it been used often?
Is there an issue tracker or anyway to give some feedback on this?
On top of my mind, some feedback:
A container is usually used to run a service, not to “install” it.
The configuration should be done via env vars.
The network setup should be done automatically so that the container is accessible from the host.
The goal cli should be available outside of the container, to access the node.
The default install should give default working params.
There should be a shared volume that saves the state of the service
The docker images we’ve generated are not supported and not intended to be used for any long-term production. At some point, we may invest the time into creating one that more closely matches your expectations, but there is nothing planned right now. I would encourage you, since you seem to have experience here, to provide us with a dockerfile to generate the image required. I would be happy to send you the dockerfile we use - it just uses the publicly-available install bootstrapper.
For now you can create a GitHub issue in the go-algorand-doc repo.
Thank you for your feedback - this helps us know what to prioritize in the future.
I just finished one that kind of work for now. It’s not as perfect as I wish, because the way shell scripts and clients are done, but it’s a good first shot. It’s probably close to what you have now as well as a bootstraper.
I’ll contribute back anyway.