10.1.2
Mac and Windows
./azcopy copy "https://REDACTED.dfs.core.windows.net/REDACTED/New%20Folder/" "/Users/me/Downloads" --overwrite=false --recursive --from-to=BlobFSLocal --check-md5 "FailIfDifferent"
(was using OAuth for auth)
The job never finished.
Create an empty folder in an ADLS Gen2 container and try to download it with a similar command.
Check if the folder is empty prior to using AzCopy.
Will add this to our backlog. Can't offer a firm ETA right now, so there's likely to a period of time when you need to use the workaround, I'm sorry.
Sounds good. Thanks John!
I actually just discovered the cause of this incidentally. For some odd reason, we infer this is a blob URL. Not sure why this behavior occurs, but I'll note it on the backlog item.
Uh. Correction, I need to investigate this more.
This one's dead simple. We missed a length check for directories. Container copies do if len(e.Transfers) == 0 and exit if the length is 0. To keep things DRY, I'll move this check to copyDownloadBlobFSEnumerator.dispatchFinalPart.
Though, considering we are about to enter a point where we refactor our copy enumerators, I'll hold off on making this change for now.
TODO: see if we fixed this in the enumeration work. I suspect we did.
Yup, this one is fixed in the enumeration work @JohnRusk. This got fixed due to our changes in STE where a transfer will error out if nothing is scheduled.
This has been fixed in the latest release-- This was caused by us not questioning the number of transfers (both in the FE and in the STE).
Enjoy!
Most helpful comment
This one's dead simple. We missed a length check for directories. Container copies do
if len(e.Transfers) == 0and exit if the length is 0. To keep things DRY, I'll move this check tocopyDownloadBlobFSEnumerator.dispatchFinalPart.