collectfast icon indicating copy to clipboard operation
collectfast copied to clipboard

AWS Multipart Uploads Breaking Hash Comparisons

Open sean-wallace opened this issue 5 years ago • 1 comments

I am having an issue where larger files in my project are being re-uploaded to AWS every single time I run collectstatic with collectfast installed. Upon inspection, this problem seems to only occur on the files whose ETags have the form XXXXXXXXXX-Y where Y indicates the number of parts in the upload. I think this may be a known issue, so wanted to ask here if there is a known solution. If not, I may be able to help tackle a fix. Thanks!

I'm currently using Python 3.8.3, Django 2.2.13, django-storages 1.9.1, and CollectFast 2.2.0

See also:

  • https://www.savjee.be/2015/10/Verifying-Amazon-S3-multi-part-uploads-with-ETag-hash/
  • https://stackoverflow.com/a/58239738/458326

sean-wallace avatar Jun 30 '20 23:06 sean-wallace

This is not a known issue to me, do feel free to address this. collectfast has "live" tests uploading to real buckets, so perhaps those could be adjusted to also include a case that will use multipart uploads.

Thanks for reporting this!

antonagestam avatar Jul 01 '20 08:07 antonagestam