[onedrive]
type = onedrive
token = {"access_token":"E=","token_type":"Bearer","refresh_token":"M"}
drive_id = 5927cd7a20529a1b
drive_type = personal
[alias]
type = alias
remote = /onedrive_local/
LOG
2021/10/27 19:10:45 DEBUG : T/VMachine/w2.tgz: Reopening on read failure after 0 bytes: retry 2/10: read /root/OneDrive/T/VMachine/w2.tgz: input/output error
It looks like a copy from local (/root/OneDrive/) to local (/onedrive_local/).
Do you also have "rclone mount onedrive: /root/OneDrive" running?
If so, then why not do the copy directly: "rclone copy onedrive: /onedrive_local/" ?
I recommend you omit flags unless they are really needed and tested for the use case. Defaults are best for OneDrive in most situations. This also goes for --transfers (and clientID).
Thanks for the quick feedback. Now with your suggestions it worked.
I mixed up by mounting OneDrive with Rclone and then using 'copy' on top.
I unmounted any rclone drives with 'killall -9 rclone'
I then updated to the lastest version with 'curl https://rclone.org/install.sh | sudo bash'
Then I confirmed the version with 'rclone version'
After I confirmed my config file was still there with 'rclone config show'
Then I started the process again using rclone -vv copy onedrive: '/onedrive_local/' and it works now.
No, I don't see a (significant) gain in my tests and therefore don't use it myself anymore.
I haven't seen a personal ClientID solve a slow/stuck OneDrive yet. It is my experience that most of these issues are due to simple misunderstandings or somebody shooting themselves in the food by adding parameters to increase --checkers, --transfers or similar.