failed to create end point from sagemaker

[Deepak-mk]

Failed to extract model data archive for container "container_1" from URL "s3://sagemaker-us-east-1-973238313667/pytorch-inference-2021-09-24-10-00-56-932/model.tar.gz". Please ensure that the object located at the URL is a valid tar.gz archive.

[enricorotundo]

I changed my setup a number of times and eventually the error message (from the endpoint status) switched to "tar.gz too big not enough memory". So maybe try doing that?

[mirodrr]

Could you elaborate on "changed my setup a number of times", and what you did to solve the "not enough memory" problem? @enricorotundo

[aasthavar]

@enricorotundo It would help a lot if you can share more details of the changes you tried ?

[enricorotundo]

FWIW I've not tracked every single change to my setup so I can't list them, sorry for that, but here's what I've tried:

• trying different compression levels with gzip, or without compression at all
• making sure no symlinks were present in the tar file
• making sure the tar file was not corrupted

Lesson learned is.. just try with a bigger instance type :)

[jkuypers93]

If anyone runs into this issue in the future, a comment on Stackoverflow led me to the correct solution: I resolved the error by giving my Sagemaker role S3 permissions (GetObject, etc.) and kms:Decrypt permissions (my S3 bucket is encrypted using KMS).

The error message does not point you in the right direction in this case