-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AssertionError on running tasks in parallel #7656
Comments
We are seeing something very similar, but also on ConcurrentTaskRunner flows. We are running 2.6.9 python 3.9 on both the agent and orion, running from the Helm chart on EKS.
|
I'm using SequentialTaskRunner and no async tasks (i.e. not running anything in parallel), and getting the same error as @deepanshu-zluri. Also running 2.6.9, but on python 3.10. We're using Prefect Cloud.
|
Hi! I believe this error is from the most recent httpcore/httpx releases — I'd recommend downgrading those dependencies to the previous version. @padbk that looks like an unrelated issue. |
Looks like they've addressed this upstream encode/httpcore#627 |
Seems that this was fixed with v0.16.2. My flows return to stable again. |
@sti0 how did you change the httpcore version ? if you are using any particular prefect image for agent can you please share it ? i added httpcore == 0.16.2 in my requirements but im still getting the same issue |
I also used that httpcore version and am getting the same issues. |
@deepanshu-zluri I'm not using the prefect base image. I build my own one and since httpcore v.0.16.2 the errors are gone away. Install prefect with pip on the build process (based on python-3.10). |
This would force install of v0.16.2 on the base image:
|
@sti0 yes i did that. still didnt work |
I have this as my image
It does not work. |
@bennnym as you are building on your own you may try to use Python base image and install prefect with pip. This is what I do and what works in my case. But maybe you run on another infrastructure or something so this won't work for you. |
Yeah I build on a custom image. That would work too, I don't see why though, I can try and report back. |
Ah, I see you both using Python 3.9? I'm on 3.10, maybe theres some difference?! And @bennnym you may like to update prefect to the latest version. Otherwise we diff apples with oranges. My setup is latest prefect 2.6.9 with 3.10. Maybe there are other interferences on the older versions. |
@sti0 is it possible to share the custom dockerfile you are using (maybe without the libraries specific to your usecase) |
Hi @deepanshu-zluri , I can't share my file directly but there is no real magic behind. Attention we pre-bundle all our stuff into one image and use it as infrastructure. Our agent and server are running in Docker images with default prefect image and Python 3.10. So it depends how you use it later. So just a little snippet:
|
Looks like our error has been solved by upgrading to 2.7.0 |
@sti0 for agent and server which default image of prefect are you using ? |
@deepanshu-zluri prefect:2.7.1-python3.10 |
did you run into any namespace issues with it ? kubernetes.client.exceptions.ApiException: (403) im getting this as soon as i changed my image. if i revert to prefect 2.6.0 on python 3.9 default image this doesnt occur |
No issues. But we are running plane Docker containers. Not using k8s |
i dont think that should make any difference |
First check
Bug summary
Reproduction
Error
Versions
Additional context
No response
The text was updated successfully, but these errors were encountered: