TejasRGitHub
commented
6 days ago Able to reproduce the same issue and found the bug.
While calculating the name following is used
f"{slugify(self.resource_prefix + '-' + self.target_label[:**(max_length - len(self.resource_prefix + self.target_uri)**)] + suffix, regex_pattern=fr'{regex}', separator=separator, lowercase=True)}"
The highlighted part truncates the bucket name to how many characters so as to accomodate the suffix ( which is the targetURI = the datasetUri in this case ). The calculation does miss 2 characters - the '-' used in between the naming.
To correct , following should be used , f"{slugify(self.resource_prefix + '-' + self.target_label[:(max_length - len(self.resource_prefix + self.suffix) - 1)] + suffix, regex_pattern=fr'{regex}', separator=separator, lowercase=True)}"
One '-' is used in between self.resource_prefix + '-' + self.target_label... and another '-' is used in the suffix which is formed of -{targetUri}.
Things to consider before making this change
Changing this logic will likely affect all the dataset iam role names and not just those but whereever NamingConventionService is used and i.e. for Environments and other places where Stacks are used. Also this same is used for generating policies.
Describe the bug
My bucket name is 50 characters long. When this bucket is imported into data.all it tries to create a dataset admin role with format: dataall-TRUNCATED_BUCKET_NAME-30ydkjf9
The problem is that this truncation is not long enough. Data.all truncated my bucket name to 49 characters. Let's add them all together
dataall-(8)TRUNCATED_BUCKET_NAME(49)-30ydkjf9(9) = 8 + 49 + 9 = 66
However AWS IAM allows only 64 characters maximum. Therefore I received an error:
How to Reproduce
Import a dataset with a bucket name longer than 50 characters.
Expected behavior
No response
Your project
No response
Screenshots
No response
OS
N/A
Python version
N/A
AWS data.all version
2.6
Additional context
No response