How to efficiently move DIR1/DIR1/* to DIR1/* (Amazon Cloud Drive, Encrypted)

Subject says it all. This is on an encrypted remote on top of Amazon Cloud Drive.

The “natural” answer would be:

rclone rmdir ACDE:DIR1TEMP

But I know first hand that, at least on an encrypted remote on top of Amazon Cloud Drive, the “move” command actually causes rclone to download the whole dir and then reupload with the new name. In the above case, there are two “rclone move” commands… so this will cause the whole dir to be downloaded and then reuploaded… twice :frowning: Faster to just purge DIR1 and reupload it again.

Anyway, is there a way around this? DIR1 has about 500GB now… :confounded:



This will be true until this issue is fixed:

If you move DIR1 up a level in the web interface then rclone crypt shouldn’t complain (assuming the web interface can actually do that - it is pretty basic!).

Hello Nick,

Great idea, using the web interface; I checked and it’s actually able to move a directory (along with everything in it) to another directory, including to the “root” directory (which the web interface calls the “All files” directory).

So, here’s the procedure I used to fix my issue, in case anyone else ends up with the same problem:

  1. used ‘rclone -v ls ACDE:DIR1’ to determine the encrypted name for DIR1:

2016/10/10 04:38:18 rclone: Version “v1.33” starting with parameters [“rclone” “-v” “ls” “ACDE:DIR1”]
2016/10/10 04:38:26 Encrypted amazon drive root ‘oettrqncmt7hfhunhkrkih7bjs’: Modify window not supported

In the above output, “oettrqncmt7hfhunhkrkih7bjs” is the encrypted name we’re after.

  1. used the ACD web interface to rename the “oettrqncmt7hfhunhkrkih7bjs” directory under the “All files” directory to some other name, say “TRASH1”.

  2. Then, I clicked on TRASH1 and found the “other” DIR1 inside it (by its encrypted name, which is also “oettrqncmt7hfhunhkrkih7bjs”), selected it, and clicked on “Move” (on the “menu bar” at the bottom of the screen); then, in the directory selection window that opened, I selected the “All files” top directory (which, again, is how the ACD web interface calls the root of the drive) and clicked on the “Move to All files” button at the bottom).

  3. Finally, still on the ACD web interface, I clicked back to the “All files” top directory and there I saw the “oettrqncmt7hfhunhkrkih7bjs” directory, and the (now empty) “TRASH1” folder. I clicked on this one and then on “Delete”.

  4. Finally, back to the command line: used “rclone lsd ACDE:” and “rclone ls ACDE:DIR1” to check that indeed the DIR1 now is under ACDE with all its files and subdirs.



1 Like

Nice one. Thanks for taking the time to make a step by step guide.

Hi Nick,

Nice one. Thanks for taking the time to make a step by step guide.

You (and rclone) are most welcome. Also, I’ve changed this topic’s title and category to indicate it’s a solved issue (or “howto”) and possibly make it easier to find for folks looking for the answer.



I had a wrong series folder uploaded to ACD
Therefore I added the correct series folder
Encoded the ACD folder name
And moved per ACD web interface the “series” folders into the new one

in the .acd folder I can see the encrypted content on my server
but within acd where it should show the “real”/decoded name the series folder is empty…

how can I refresh the mount to show me the content?
as the web interface does?

Hello Alneven,

how can I refresh the mount to show me the content? as the web interface does?

What I do is to stop the mount (kill the rclone mount process, and fusermount -u its mountpoint) then bring it up again.

I don’t currently know if there’s a less “windowsy” (ie, not bringing it all down to start all over again) way to do it, perhaps someone else (@ncw ?) will chime in?



oh, thanks

I just did the fusermount and nothing got refreshed
I will test the rclone kill and the fusermount later this day

Thanks for the idea

Hello @alneven,

Uh? So you ran “fusermount -u” onto the mountpoint without stopping “rclone mount” first?
You probably just got a fat useless error message (as the mountpoint would have been busy and so the system would refuse to umount it). No wonder nothing got refreshed…

You are supposed to stop “rclone mount” first, then run “fusermount -u” on the mountpoint. This should “refresh” everything, as it’s a new rclone process that will be talking to the remote and reading in all current metadata (directory contents, etc).

Please let us know how it goes.