Connection test just shows red rectangle #26048
Replies: 3 comments 8 replies
-
|
I've been having signal 11-related problems lately (ever since I started messing around with providers). Probably the core of this has something to do with a provider and is not a core airflow bug. Still, I think we should be a little more verbose when it comes to communicating this error. |
Beta Was this translation helpful? Give feedback.
-
|
you mean, as opposed to a red rectangle? |
Beta Was this translation helpful? Give feedback.
-
Surely. Signal 11 is SIGSEGV which generally means that there is a serious error with your installed libraries. I.e. we have completely no idea what is the error because likely your deployment is messed up (for example you have broken libraries installed, wrong binaries in your system, badly configured libc* librarires or similar). It's generally rather bad error indicating that you have done something, airflow maintainers could not foresee and generally it crashes python interpreter, so in-general there is no way to provide any information, because at the moment it crashes, we have no way to handle it. But If you think there is a better way and you would like to submit a fix to make a bettter message in this case - we are absolutely open to accept a fix. Since we do not know what is the root cause (somewhat broken environment in the way that crashes C-level library) you are probably one of the best people to be able to provide the fix - because you have the unique environment that generates this error. And since Airflow is created by > 2100 contributors - people like you who often contributed a fix when they had problem, you are most welcome to join the team. I will convert it to discussion and if you will be able to provide a fix that will provide a better description, that would be fantastic. |
Beta Was this translation helpful? Give feedback.
-
|
BTW. I recommend you to start from a clean state or restoring your OS to a clean state if you want to have Airflow working (ideally after you work out and submit the fix above to make a better description of the problem). I think there is no other easy recovery if your low-level c-libraries are messed up. |
Beta Was this translation helpful? Give feedback.
-
|
I also have this issue. We are on version Version: v2.2.3 |
Beta Was this translation helpful? Give feedback.
-
|
I did a export no_proxy=* in the web pod and still get the same error. or do you mean in the airflow.cfg file? |
Beta Was this translation helpful? Give feedback.
-
|
Oh if you run in k8s that mean known macOS issues with fork process not affect to you. Is your error somehow related to signal |
Beta Was this translation helpful? Give feedback.
-
|
And if you also find in logs
|
Beta Was this translation helpful? Give feedback.
-
|
Hey @Taragolis This is the traceback for me: |
Beta Was this translation helpful? Give feedback.
-
|
Yes. Segmentation fault is the sign that this is MacOS ARM problem. Similar Issues https://github.com/apache/airflow/discussions?discussions_q=Negsignal.SIGSEGV -> you will find more details and possible workarounds there. Setting environment variable NO_PROXY='*' helps. You just have to make sure it is set and exported in your POD. You likely will need to build your own image to set it, it's not enough that you enter the container and set it - it has to be set for the whole image. One of the ways to do that is to extend the image. You can see more description about buildin your image and extending the base one here https://airflow.apache.org/docs/docker-stack/build.html#examples-of-image-extending (you likely need to use ENV command). And then you need to use the image. Let us know if it worked for you. |
Beta Was this translation helpful? Give feedback.
-
Apache Airflow version
main (development)
What happened
I was walking through this tutorial: https://astro-sdk-python.readthedocs.io/en/stable/getting-started/GETTING_STARTED.html which had me create an aws connection. After doing so, I clicked the "test" button, and saw this:
And this in the webserver logs:
What you think should happen instead
I'm assuming that the error message was empty because the worker which was supposed to populate it died before it got a chance. If this happens, I'd still like to get something in the UI. For instance, it could say:
Worker with pid 64227 was terminated due to signal 11which would at least give me something.How to reproduce
test a connection. I've also seen this with the snowflake provider, so I don't think it's aws specific
Operating System
Mac OS
Versions of Apache Airflow Providers
apache-airflow-providers-amazon==5.0.0
apache-airflow-providers-common-sql==1.1.0
apache-airflow-providers-ftp==3.1.0
apache-airflow-providers-http==4.0.0
apache-airflow-providers-imap==3.0.0
apache-airflow-providers-snowflake==3.2.0
apache-airflow-providers-sqlite==3.2.0
Deployment
Virtualenv installation
Deployment details
I've been setting up airflow venvs like this (nothing especially weird in this case)
python scripts/ci/pre_commit/pre_commit_compile_www_assets.pyairflow webserverAnything else
No response
Are you willing to submit PR?
Code of Conduct
Beta Was this translation helpful? Give feedback.
All reactions