[s3/windows] HeadObject operation: Not Found

[prashaw]

Please provide information about your setup
DVC version(i.e. dvc --version), Platform and method of installation (pip, homebrew, pkg Mac, exe (Windows), DEB(Linux), RPM(Linux))

Hello,

I get this error when using dvc add on s3 bucket when the files are deep inside the directories.

C:\Users\user\Desktop\AD>dvc add s3://samplebucket/dvc-test/6b –recursive
Adding...
ERROR: s3://samplebucket/dvc-test/6b/test\test\test.csv does not exist: An error occurred (404) when calling the HeadObject operation: Not Found

The error visible here is with mismatched nt vs posix paths. This works without issues in ubuntu 18.04.

DVC version:

C:\Users\user\Desktop\AD>dvc version
WARNING: Unable to detect supported link types, as cache directory '.dvc\cache' doesn't exist. It is usually auto-created by commands such as `dvc add/fetch/pull/run/import`, but you could create it manually to enable this check.
DVC version: 0.93.0
Python version: 3.7.4
Platform: Windows-10-10.0.17763-SP0
Binary: False
Package: pip
Supported remotes: azure, gdrive, gs, hdfs, http, https, s3, ssh, oss
Repo: dvc, git

The commands I followed prior to this including setting correct aws credentials are:

pip install awscli, boto3, dvc[all]
git init
dvc init
git commit .dvc/config -m "Configure bitbucket remote"
dvc remote add s3cache s3://samplebucket/dvc-test/d4 -f
dvc config cache.s3 s3cache

Thank you!

[xmaayy]

I am also getting this bug, any solutions?

[efiop]

Hi @xmaayy ! Are you also using cache.s3? If so, you are probably misusing --external. What you really want is #4520 , which we don't have implemented yet, so you need to download the data manually to your local machine, dvc add it and dvc push it.

[efiop]

Closing in favor of #3920 for the scenario itself