Undersanding the --inplace command

I am running the following command:

rclone sync --progress --transfers=1 gdrive: koofr:'gdrive backup'

It is working fairly well, but I still get too many 500 errors. Adding --transfers=1 helped reduce the number of errors considerably, but I was looking for other ways to tweak the command and came across --inplace. I read the documentation, but am not very clear if this would be appropriate to use in this scenario and whether or not it might help?

(Sorry, I'm not posting full logs and bug reporting information because I don't see this as a rclone issue so much as limitations with the APIs of Google and Koofr that I'm trying to workaround. It is mostly giving me errors with very large files, but not always. Adding --transfers=1 greatly reduced the error count and I'm trying to look for further tweaks I could try.)

Current failure rate is about 17%. Before I switched to using --transfers=1 it was more like 80-90%. I'd like to get it down to below 5% if possible, but I'm not sure what else can be done?

--inplace only applies to local and (s)ftp. See docs for details.

Try to limit number of transactions per second, try to disable HTTP2 and maybe only bind to IPv4. These are classics of problematic connection performance tricks. As you do not post any details so impossible to guess what the problem is.

rclone sync gdrive: koofr:'gdrive backup' --progress --transfers 1 --checkers 1 --tpslimit 1 --tpslimit-burst 0 --disable-http2 --bind 0.0.0.0

On the other hand sometimes errors are unavoidable. Rclone usually will retry.

1 Like

Thanks a lot! This seems to be making things more reliable without hurting my speed, which is what I wanted.

You could to experiment a bit by increasing some values. Hard to tell what is the sweet spot for your particular setup.

1 Like

These settings seem to have doubled my speed (even compared to my initial settings) without increasing the error rate (at least not so far):

rclone sync gdrive-media: koofr:'fulaoshi media' --progress --transfers=4 --checkers=4 --tpslimit=5 --tpslimit-burst 10 --drive-chunk-size=256M --disable-http2 --bind 0.0.0.0

Possibly I could improve speeds even more, but this is good enough for now. It should be able to sync 375GB in 21 hours if it keeps on at the current pace. (Before it was 2 days, 11 hours.) Thanks again.

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.