Can Sagemaker RStudio mount an NFS share?

Question

Our existing Python code currently accesses data files residing on a file system. It navigates through folders to reach individual files and reads them.
 
We would now like to migrate the code to notebooks hosted in Sagemaker Studio. The data files will be hosted in S3 buckets.
However, we do not want to modify the code to use S3 API to access files. Instead, we would like to continue using a filesystem view of the files. What is the best way to make the S3 objects visible to the Sagemaker Studio as a filesystem?
Please note that we would like to mount in Sagemaker studio and not on Sagemaker notebook.

Answer

SageMaker Studio uses EFS for storage. So you could potentially do an `aws s3 sync` and sync your S3 files/folders to EFS and use them as a filesystem. I'd also recommend to be cautious of the file sizes since EFS is more expensive.

Studio now supports JupyterLab3 and you can now install popular extensions. So if you're just looking at *viewing* the S3 files from Studio, you can use the [S3 file browser extension](https://github.com/IBM/jupyterlab-s3-browser).

Can Sagemaker RStudio mount an NFS share?

関連するコンテンツ