https://prefect.io
Join Slack
Got another issue with new deployments, it seems that manifest_path is still required to have a valu...
# ask-community
o
Got another issue with new deployments, it seems that manifest_path is still required to have a value in the deployment.yaml file. Is this intended?
Copy code
pydantic.error_wrappers.ValidationError: 1 validation error for Deployment
manifest_path
  none is not an allowed value (type=type_error.none.not_allowed)
✅ 1
a
Do you use venv? I'd recommend trying with a new fresh conda environment and installing Prefect there
o
I do use venv! I re-created it when I upgraded to 2.0.4
Although I don't use conda
a
again I recommend conda or just not venv, your call to make
I saw one person had similar issue
to troubleshoot and see that venv is at fault by using the old Prefect version, try replacing the default command in YAML with:
Copy code
command:
    - python
    - -c
    - 'import prefect; print(prefect.__version__)'
most likely it will show you the old Prefect verison
o
Giving it a go..
It printed "2.0.4"
I realized this has to do with where I run the agent by the way, that's just a VM that doesn't use venv. Re-created the VM two hours ago and ran pip install prefect==2.0.4, that's about it
I've worked around the issue by pointing it at an old manifest file I had lying around, that I've uploaded to blob storage to make it happy. Probably shouldn't be necessary though. 🙂
🙌 1
d
I was facing the same issue. I was using docker image stored on cloud for the agent. I had not updated the docker image with the new version of prefect. The workflows ran successfully after I rebuilt the image and pushed it to dockerhub. Just posting this in case some one else has the same issue.
🙏 1
o
Yeah, that explains my issue.. My bad. Thanks Bigya (and Anna)!
🎉 2
a
Nice work everyone and thanks for sharing both problems and solutions here, very helpful!
5 Views
Open in Slack
PreviousNext
Powered by

Prefect Community

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.

Powered by