OneDrive for Business - always upload for a while and then accessDenied


#1

rclone v1.41

  • os/arch: linux/arm
  • go version: go1.10
    OneDrive for Business E3 developer (MSDN)

I try to use rclone on linux. When I use rclone copy or rclone sync, it always upload a while (4-5 minutes) and then shows
2018/05/03 23:11:38 ERROR : Sharpay’s.Fabulous.Adventure.2011.HR-HDTV.x264.AC3.Chs.Eng-Deefun.mkv: Failed to copy: accessDenied: The caller does not have permission to perform the action.

Every big file stoped because of this accessDenied. I really don’t know how to fix it…


#2

Does this happen to random files or specifically this file. Try renaming this file and see in cases I know special characters like % and some others are not permitted.

What is the file size.


#3

Thanks for replying,
Now this problem is fixed by doing nothing. But there is another problem.

activityLimitReached: throttledRequest: The request has been throttled

This is like the last problem. The file upload a while (this time longer) and then appears this error.
This error appears like every day at the same time. My file size is around 250MB-400MB. There is a lot of them. And it happen to every file which is uploading. Every day at this time, none of my file will transfer successfully. But later after many hours, it will be fixed by doing nothing.


#4

This happens with me as well.

I think MS has some daily cap or something like that. Either after sometime the upload speed will reduce or I see this message.

I wait for 2 or 3 days and start again than it runs longer


#5

According to my observation, this error happen from 18:30 PST to 3:30 PST like every day. Now it is not throttled. It seems that MS has some limit every day.


#6

Hi,

this error happen from 18:30 PST to 3:30 PST

I can support this observation. Last Friday I got very good transfer rates (15 simultaneous transfers, ~10 MB/s overall) in the afternoon. Now this Monday morning before 10h30 UTC the rate was extremely poor (few kB/s with 5 transfers, many throttled requests), but it seems to have “lifted off” at around 10h45 UTC. I am now (11h UTC) getting very few “activityLimitReached: throttledRequest” notices.

Note: I don’t use the --user-agent flag nor an own AppId. I may not need to try it if sending big files in the UTC afternoon works reliably.

This said, I observe something a bit odd: in the progress block, for each file the ratio of transferred data for each file quickly rises up to 100%, then the file transfer rate drops, but my network load monitor still shows high outgoing rate. E.g. all 5 currently transferred files are at 100%, show a file transfer rate around 10-20 kB/s, totalling ~100 kB/s, but the monitor says ~1-2 MB/s, which is a substantial difference! (The monitor does display MB/s and not Mb/s).


#7

Likely you are seeing the effect of internal buffering of chunked uploads. There is a 10MB buffer for chunked uploads by default and the bandwidth monitor will show the transfer as the file is read into the buffer, not as it is sent from the buffer.

It is possible to fix this - if you please make a new issue on github we can have a look at it!