You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Have you tried using the environment variable? Maybe is just a matter of updating the documentation to reflect that it can be defined?
I have no experience with this part of the codebase but just for validation purposes.
raulcd
changed the title
Support AWS_ROLE_ARN env variable for pyarrow.fs.S3FileSystem input
[Python][C++] Support AWS_ROLE_ARN env variable for pyarrow.fs.S3FileSystem input
Mar 7, 2025
yes, I've tested this, attempting to use role results in an error of the following form
OSError: When getting information for key 'schemas/meta.db/trino_queries_iceberg/metadata/00000-41568416-bc76-4236-afab-a7bec772eb32.metadata.json' in bucket 'REDACTED-BUCKET': AWS Error ACCESS_DENIED during HeadObject operation: No response body.
the following is a workaround for this, but probably not great as it leads to a static credential for a short lived credential.
So, if I understand correctly, using the AWS_ROLE_ARN environment variable works but there seems to be an issue on the AWS SDK where it seems to use an anonymous user when using AWS role arn. This is both using the environment variable or using the parameter role_arn from our API , as described on the issue linked, right?
Describe the enhancement requested
Documentation here https://arrow.apache.org/docs/python/generated/pyarrow.fs.S3FileSystem.html
mentions supporting AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and AWS_SESSION_TOKEN but strangly role id can only be taken as role_id argument
request to support AWS_ROLE_ARN env variable instead
Component(s)
Python
The text was updated successfully, but these errors were encountered: