OverwriteNewerLocalError raised despite force_overwrite_from_cloud=True in parallel read only scenario

[YishGene]

Hi,

In v0.20.0, I'm running into an issue where, when trying to read an S3 file from multiple readers in a parallel scenario (multiple readers, no write access), I am passing force_overwrite_from_cloud as instructed in the error report, but still getting an OverwriteNewerLocalError.

I will try to set different caches for each process, but this could be wasteful as I am only reading and will have to copy already downloaded files.

Related to isses: #283 and #128 , but not exactly.

The error reads like this:

...
fp = file_path.open('rb', force_overwrite_from_cloud=True)
File "/opt/conda/lib/python3.10/site-packages/cloudpathlib/cloudpath.py", line 686, in open
self._refresh_cache(force_overwrite_from_cloud=force_overwrite_from_cloud)
File "/opt/conda/lib/python3.10/site-packages/cloudpathlib/cloudpath.py", line 1303, in _refresh_cache
raise OverwriteNewerLocalError(
cloudpathlib.exceptions.OverwriteNewerLocalError: Local file (/tmp/tmp9h3c_4qk/s3://|bucket|/|ObjectPath|/GT_32.h5) for cloud path (s3://|bucket|/|ObjectPath|/GT_32.h5) is newer on disk, but is being requested for download from cloud. Either (1) push your changes to the cloud, (2) r:emove the local file, or (3) pass force_overwrite_from_cloud=True to overwrite; or set env var CLOUDPATHLIB_FORCE_OVERWRITE_FROM_CLOUD=1.

[pjbull]

Thanks for this report. I'm going to close this issue since it is almost certainly the same root cause as #283.

Let's continue the conversation there. I will add some thoughts.