how do transfers of large files work?

Hi, While I was testing sync thing I tried copying a large file to test the performance. Initially I set syncthing up on my Desktop and on a raspberry pi. I transferred the file, which went quite slowly. I am lead to believe however that this may be due to the fact I had the web GUI open. Later I setup a 2nd raspberry pi( I happen to have a stack of them). I ran the same sync again and watched the process through the GUI.

I thought that syncthing transferred like a torrent in that it broke the file into pieces and transferred them simultaneously between multiple clients. Can it only read pieces from complete files of nodes that have already synced? When I ran the transfer with 2 RPis the files first transferred to one RPi and then the transfer process started again on the other.

Just keen to know how things work.



Since 0.13 peers that do not have the whole file yet can also participate in providing data, yet the announcements about the data available get sent every 5 seconds and only apply to files of atleast a few mb.

Since version 0.13, Syncthing transfers incomplete files. Before that, the client had to have the complete file to be able to announce, that it has that file.

Having the gui open or not should not make a significant difference in current versions.

quite slow

Define “quite slow”. As the RPi isn’t that powerful, it won’t be able to transfer as fast because of the encryption.

Did you connect the desktop to both RPis?

I define quite slow as 485KBps (I assume bits?) I know there is also encryption going on. Owncloud uploads at around 4.0MB/s (megabytes). I know the RPI isn’t so powerful, but I was expecting a bit more.

Yes I connected all the devices together. I was expecting that the sync would simultaneously run to both RPI1 and RPI2.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.