The beauty of docker is that it is a reflection of how much someone cares about deployments: do you care about being efficient? You can use `scratch` or `X-alpine`. Do you simply not care and just want things to work? Always go for `ubuntu` and you're good to go!
You can have a full and extensive api backend in golang, having a total image size of 5-6MB.
I've done both, tiny scratch based images with a single go binary to full fat ubuntu based things.
What is killing me at the moment is deploying Docker based AI applications.
The CUDA base images come in at several GB to start with, then typically a whole host of python dependencies will be added with things like pytorch adding almost a GB of binaries.
Typically the application code is tiny as it's usually just python, but then you have the ML model itself. These can be many GB too, so you need to decide whether to add it to the image or mount it as a volume, regardless it needs to make it's way onto the deployment target.
I'm currently delivering double digit GB docker images to different parts of my organisation which raises eyebrows. I'm not sure a way around it though, it's less a docker problem and more an AI / CUDA issue.
Docker fits current workflows but I can't help feeling having custom VM images for this type of thing would be more efficient.
> You can have a full and extensive api backend in golang, having a total image size of 5-6MB.
So people are building docker "binaries", that depend on docker installed on the host, to run a container inside a container on the host– or even better, on a non-linux host, all of that then runs in a VM on the host... just... to run a golang application that is... already compiled to a binary?
xandrius|9 months ago
You can have a full and extensive api backend in golang, having a total image size of 5-6MB.
hereonout2|9 months ago
What is killing me at the moment is deploying Docker based AI applications.
The CUDA base images come in at several GB to start with, then typically a whole host of python dependencies will be added with things like pytorch adding almost a GB of binaries.
Typically the application code is tiny as it's usually just python, but then you have the ML model itself. These can be many GB too, so you need to decide whether to add it to the image or mount it as a volume, regardless it needs to make it's way onto the deployment target.
I'm currently delivering double digit GB docker images to different parts of my organisation which raises eyebrows. I'm not sure a way around it though, it's less a docker problem and more an AI / CUDA issue.
Docker fits current workflows but I can't help feeling having custom VM images for this type of thing would be more efficient.
endofreach|9 months ago
So people are building docker "binaries", that depend on docker installed on the host, to run a container inside a container on the host– or even better, on a non-linux host, all of that then runs in a VM on the host... just... to run a golang application that is... already compiled to a binary?
anthk|9 months ago