Hmm, that is showing 24 sockets open to the remote. Can you share your command line please? Rclone should only use --transfers + --checkers sockets normally.
Are you uploading big files BTW? I wonder if the chunked upload is blowing up the concurrency limit…
There is also a suspicous number of sockets open in lsof which looks strange to me, though these aren’t matched by the number in netstat.
I can see from the lsof that you are using --transfers 4. You are probably using --checkers 8 which is the default so there should be no more than 12 sockets open. However there are 24 open which is odd. Also 131 “protocol: TCP” which I don’t really understand.
Can you do a run of rclone then kill -QUIT it in the middle of the run - either use kill -QUIT or press CTRL-\. This will produce a dump of everything that rclone is doing - can you paste that!
Hi . i am still trying figure whats the problem… ( i am not super user so hard to figure the problem)
here is is the Code you asked for killed rcloned in the middle of the run using CTRL-\ hope this will be helpfull https://pastebin.ubuntu.com/p/S7hkP8QJ5g/
I just updated to 1.41 but I still get an upload loop which possibly relates to this bug. I’m using a OneDrive for Business account. rclone runs on Mac OS 10.13.4.
I have four bigger files which are being uploaded but after about 1 h 40 min the upload fails and the upload starts again.
Your sync appeared to retry - which is fine - stuff happens on the internet. The error I can see in the log has this message read tcp 192.168.0.87:61576->13.107.6.151:443: read: connection reset by peer which means that something closed the connection - maybe onedrive, maybe your ISP, maybe your router.
The retry didn’t work properly because of this error
2018/05/10 20:07:54 ERROR : rclone_backup_sync_logfile.txt: Failed to copy: can’t copy - source file is being updated (size changed from 411 to 2152)
So you are syncing the log file of this sync by the look of it… I suggest you --ignore that file otherwise your syncs will never complete properly.
Thanks a lot for looking into this, Nick! I’m now trying your suggestion will update you on the outcome. I don’t understand why the error regarding the log file would relate to the continuous interruption of the other file uploads but if the exclusion of the log file would help I’d be more than happy of course. If it doesn’t help I’m afraid that OneDrive for Business may have a upload timeout in place if the upload of a file takes too long. But there’s also a chance that the issue could relate to my ISP (internet via TV cable) since I found a warning message in the router status regarding the WAN connection. If the issue persists I will first look into this potential issue with my ISP. Thanks again, Helge
It works, Nick! You were right, it was the exckusion of the log file which probably couldn’t be accessed by rclone for upload. I still don’t get how this caused the upload interruption of the other files. I guess it may be a rsync related thing. Thank you for your help! Helge