Hi 👋, I'm new to Prefect and am testing the push work pools with Google Cloud Run. By going through the
how-to guide
, a push work pool was successfully provisioned. However, running the Prefect CLI generated example_deploy_script.py results the following error message, After restarting Docker Desktop on an Apple M2 machine, the error message remains the same. Any thoughts? Many thanks!
Copy code
DockerException(docker.errors.DockerException: Error while fetching server API version: ('Connection aborted.', FileNotFoundError(2, 'No such file or directory')
j
Jeff Hale
02/23/2024, 3:57 PM
Hmm. If you upgrade your Prefect version to 2.16.0 and rerun the script, do you now see an error about Docker not running? (we updated the error message that should appear if Docker isn’t running so that it’s more descriptive).
Can you reach Docker from your CLI? e.g. with
docker run hello-world
?
w
Wayne Lai
02/24/2024, 7:22 AM
Yes,
docker run hello-world
works well. The same result remains after reboot.
j
Jeff Hale
02/24/2024, 1:42 PM
There are a few potential fixes in this Docker thread on working with Docker and Python that might be helpful
w
Wayne Lai
02/25/2024, 2:45 PM
In the Docker Desktop, I noticed that the "default" builder was down due to some error. However, another builder, "desktop-linux" has been running and that's why images can still be built. Two questions 1) can Prefect use builders other than the "default" one? and 2) can Prefect uses Google Cloud Build instead? Thanks!
j
Jeff Hale
02/25/2024, 3:00 PM
Hmm. I'm not sure. My quick look suggested the Linux builder wouldn't build an image on a Mac. Are you able to build images?
Prefect is using the docker Python package under the hood, so I don't think there's a way to use Google Cloud Build.
Alternatively, you can build your own Docker image and use it with a push work pool. Just specify it in the work pool.
w
Wayne Lai
02/25/2024, 4:03 PM
Yes, images can be built. Rosetta is installed on my machine. Thanks for help. I'll give Serverless Infrastructure with Prefect Workers a try instead.
Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.