Skip to content
This repository has been archived by the owner on Sep 2, 2024. It is now read-only.

[bug] csv_provider Docker image fails to start #119

Closed
sophokles73 opened this issue Jul 4, 2023 · 2 comments
Closed

[bug] csv_provider Docker image fails to start #119

sophokles73 opened this issue Jul 4, 2023 · 2 comments

Comments

@sophokles73
Copy link
Contributor

With 665440f the Docker image created for the csv_provider no longer starts up but fails with the following error:

> docker run ghcr.io/eclipse/kuksa.val.feeders/csv-provider:main
Traceback (most recent call last):
  File "provider.py", line 128, in <module>
  File "asyncio/runners.py", line 44, in run
  File "asyncio/base_events.py", line 649, in run_until_complete
  File "provider.py", line 79, in main
  File "kuksa_client/grpc/aio.py", line 55, in __init__
TypeError: BaseVSSClient.__init__() got an unexpected keyword argument 'tls_server_name'
[7] Failed to execute script 'provider' due to unhandled exception!

My understanding is that the BaseVSSClient's init function does not accept arbitrary kwargs at that position in kuksa-client 0.3.x.

The init function's signature seems to have been changed in 0.4.0a3 to support this. However, the pip install command used in the csv_provider's Dockerfile does not match against pre-release versions and therefore picks up the incompatible kuksa-client 0.3.1 instead.

I see two ways forward here:

  1. adjust the Dockerfile's pip install command to include --pre
  2. do not include the tls_server_name argument when constructing the VSSClient

FMPOV option 1 would be more appropriate and also seems to have been anticipated by #93.
I would be happy to provide a PR ...

@erikbosch WDYT?

@erikbosch
Copy link
Contributor

I think alternative 1 is the best. I have also written down as we as part of release should remove any --pre, see https://github.com/eclipse/kuksa.val.feeders/wiki/Release-Process

But we might have another problem, in requirements.txt we state that version shall be >0.3. Looking at the build log it seems to use a cached version. The intention of >0.3 is that it shall complain/fail if it cannot find any 0.4+ version, but that does not seem to happen. So maybe we need to change the build config as well. I did a quick test at #120 but it is stuck in the queue, will check tomorrow.

@erikbosch
Copy link
Contributor

I think this one can be closed as #120 is merged

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants