Hi folks I'm trying to setup my Prefect v2 flows to run in `KubernetesJob`'s but I'm having trouble figuring out the best way to setup my storage block. In my deployments, I create a docker image with all my flow code and corresponding dependencies. I wish I could use something like the

DockerContainer

storage block to my deployment indicating that it should just pull my docker image from AWS ECR and run the flow. However, as far as I understand, the

DockerContainer

block is also an infrastructure block and is meant to run the image against docker directly on a machine - i.e., NOT in a

KubernetesJob

. This means that so far, I've been relegated to using the

S3

storage block which pushes my flow code to S3, only to pull it back to my docker image at flow run time (when the image already contains all the code it needs to run). This feels wrong so I'm sure I must be missing something here. How do I configure

KubernetesJob

to just pull a docker image and not rely on any additional storage block like S3? PS: so far, I'm really enjoying how quickly I'm getting up to speed on prefect v2. Despite of kinks like this, I'm really enjoying the development experience.

Hi! Just set an

image

on your Kubernetes Job — although for ECR it’s a little tricky to authenticate.

I am doing that but do you mean that I don't need to specify the flow storage for that deployment?

See https://github.com/PrefectHQ/prefect-aws/pull/68 for WIP on a block for ECR authentication

I have the ECR auth down - I am running the flow in the correct image pulled from ECR. I see

Yeah, I think you might need to specify to skip upload too.

oh, how do I skip upload?

I’m not sure where our latest demo of that is tbh, I can find out tomorrow.

no worries, thanks a lot

There’s a

--skip-upload

flag on the

prefect deployment build

command

Just to confirm - this worked! I was able to bypass S3 storage and just use my docker image