Debug Python script to upload files to AWS Glacier -- 2

I have a little python script that uploads PDF files to Amazon Glacier and saves the status in a database table. ([login to view URL]) We currently use it to upload files from an S3fs directory that is actually an S3 bucket mounted on the linux filesystem ([login to view URL]). However, there is an issue. The s3fs directory contains many PDF files (1MB or more in size each). When the script is run it doesn't upload anything. It seems to get stuck on the first multi-part upload. However, if the script is run to upload files from a normal directory, it works fine. I believe this is happening because s3fs is not very good at metadata operations, like listing files and directories, especially large number of files. The script currently loops through each file in the directory. This is just a theory I need an excellent python coder to figure out what the problem is and to either: 1) Fix the issue with the script/s3fs combination so that it works as designed -or- 2) Modify how the script operates. Instead of looping through the filesystem, it should rely on the database. It needs to first select any filenames where archive id is missing from the database, and then loop through that list and explicitly upload each specific filename to Glacier, afterward updating the database as it currently does. I believe S3fs will have no problem if you operate on a specific file from its filepath. I don't know this for sure, but it's worth a try. -or- 3) Write a better script!! Coder must set up his own test environment as described above to find the solution.