Hi everyone, we just updated from 0.15.4 to 0.15.9 and suddenly our flows are failing on a specific ...

Maurits de Ruiter

11/23/2021, 4:45 PM

Hi everyone, we just updated from 0.15.4 to 0.15.9 and suddenly our flows are failing on a specific line of code. Every instance of

Copy code

if status in ['SUCCEEDED', 'FAILED']:

fails with the following error:

Copy code

TypeError: 'sequence' not supported between instances of 'str' and 'tuple'

If we log the type of string and array, it returns their types correctly.

Maurits de Ruiter

11/23/2021, 4:46 PM

It also fails with the same kind of error on:

Copy code

if 'Contents' not in page:

Where page is a dict. The error here is:

Copy code

TypeError: 'dict' not supported between instances of 'str' and 'dict'

Jake Kaplan

11/23/2021, 4:50 PM

Hi Maurits, do you have a minimum reproducible example that works with 0.15.4 and fails with 0.15.9?

Maurits de Ruiter

11/23/2021, 5:00 PM

Hi Jake, Well, a test was working locally on my pc, so I looked at the differences with the ECS Agent and it seems the Python-version was the problem. Locally I was running 3.8 and the ECS task used the prefecthq/prefect:0.15.9-python3.9 image. After changing this to the python3.8 tag, it was working again.

Jake Kaplan

11/23/2021, 5:04 PM

I am not sure off of the top of my head a change between python3.8 and 3.9 that would cause this type of error (prefect should support both). I'll try to see if I can find something. I'm glad you were able to resolve your issue for now!

Maurits de Ruiter

11/23/2021, 5:05 PM

Yeah, it’s kinda strange, because I updated locally as well, to test this issue. Locally it’s still working as expected, no sudden errors

Maurits de Ruiter

11/24/2021, 4:27 PM

Okay, so I just ran a flow locally with Python3.9, which worked fine. Then I registered it with an ECSRun on a Python3.8 container, and it failed again, now with this error: SystemError(‘unknown opcode’)

Maurits de Ruiter

11/24/2021, 4:28 PM

It seems to be a problem when you work locally in another Python version then the environment where the flow eventually runs

9 Views

Open in Slack

Previous Next

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.