Recursive copy script for odrive to copy files from could to cloud.

This script is intended to allow batch copy operations between clouds accessible to odrive, including:

• Amazon Cloud Drive
• Google Drive
• OneDrive
• S3
• B2
• Not-actually-cloud things like FTP, SFTP, and WebDAV

This script was written after Multcloud failed repeatedly to copy a few hundred gigabytes of data from Amazon Cloud Drive to Google Drive, and competing cloud migration serrvices offered to do it for preposterously high prices. It does need to download and upload each file, so if you want faster transfers or are worried about transfer quotas you may want to run it on a VPS.

In light of the recent banning of the open-source Amazon Drive clients by Amazon, the usage of odcopy.sh will be explained in the context of migrating from Amazon Cloud Drive to Google Drive. These instructions assume a Unix system, and have been developed on Ubuntu.

0. Grab the odcopy.sh script from this repository.

wget https://raw.githubusercontent.com/interfect/odrive-copy/master/odcopy.sh
chmod +x odcopy.sh

1. Open an odrive account for free. Turns out that "unsync" is a premium feature, so if you don't have a free trial you will have to somehow make that work.

2. Connect your cloud accounts: Amazon Cloud Drive and Google Drive. The names you assign to the accounts will be their paths in the folder used by odrive later.

3. Download and start the odrive agent, which is the proprietary binary odrive uploader/downloader process. It stays running as a server process.

4. Download and install on your PATH the odrive.py CLI tool, which is used to control the server.

5. Create a new Authentication Code for your odrive account.

6. Authenticate your local client with odrive.

odrive.py authenticate YOUR-AUTH-CODE

7. Make a directory for odrive to keep files in. It only needs to have as much space as your largest file (plus placeholder files). Make sure it's on a filesystem that supports long paths (i.e. not an encrypted home directory).

mkdir /tmp/odrive

8. "Mount" your odrive. This isn't a real FUSE mount; it just tells odrive to synchronize stuff using this as a root. The odrive server will create placeholder .cloudf file for the root directory of each cloud you have added on odrive. You can "sync" inividual folders to pull down their contents as placeholder .cloud files (for files) and .cloudf files (for subfolders). You can "sync" the .cloud files to pull down file contents, and you can "unsync" files and foilders to turn them back into the placeholders.

odrive.py mount /tmp/odrive /

9. Sync (non-recursively) all the directories down to the ones you want to move. If you want to copy Amazon Drive/data to Google Drive/data, you would have to sync Amazon Drive.cloudf and Google Drive.cloudf in order to pull down placeholders for their immediate contents.

odrive.py sync /tmp/odrive/Amazon\ Drive.cloudf
odrive.py sync /tmp/odrive/Google\ Drive.cloudf

10. Start the copy operation. The source folder must exist and be unsynced (i.e. be a .cloudf). The destination folder may exist, but if it does it must also be unsynced.

./odcopy.sh /tmp/odrive/Amazon\ Drive/data.cloudf /tmp/odrive/Google\ Drive/data.cloudf

The script will recurse through folders and subfolders, downloading each file from the source ansd uploading it to the destination. When it finishes, all the files should be copied across.

11. When you're done, unmount the directory and clear out all the leftover placeholder files.

odrive.py unmount /tmp/odrive
rm -Rf /tmp/odrive

The script won't start unless the folders you want to copy between are unsynced. If you abort the copy in the middle, or something goes wrong, you may need to manually unsync the folders:

odrive.py unsync /tmp/odrive/Amazon\ Drive/data.cloudf
odrive.py unsync /tmp/odrive/Google\ Drive/data.cloudf

There are many ways this script could be made more efficient or robust. Pull requests are gladly accepted.