I thought I'd stick my robocopy job up that I've been using for testing - succesfully so far - on the chance it might help someone else achieve what I've been trying to achieve .
@echo
@echo Begin Monitoring
@echo
robocopy P:\ E:\ /E /Z /XA:H /W:10 /MT:2 /MON:1
This runs on drive P:\ to drive E:\ and the switches do the following:
/E - copy subdirectories, including Empty ones.
/Z - copy files in restartable mode.
/XA:H - excludes hidden files.
/W:10 - waits 10 seconds between retries
/MT:2 - uses two threads for copying
/MON:1 - monitor the source (P:\) and run every 1 minute
Because I'm not using the /MIR (mirror) switch, when files are deleted from my Drivepool drive (P:\), the Robocopy job picks this up, but does nothing. This is exactly what I want as I can look back through the job and see if files have been deleted, and if accidental, I can then copy them back down from my cloud storage.
Here's a picture for verification of that - I've run a test by deleting (all duplicated test data of course!) a folder to simulate an accidental delete. The job runs every minute and detects the change, that the cloud drive now has all these extra files:
So I can work out from here what I need to restore.
I'll be working on the batch file so that I can output a proper log file of all folder differences to detect accidental deletes to make things easier, but the jist of this is it works!