Hi
I want to use cwrsync to synchronize files on Windows notebooks with a Windows Server. I don't like Windows but that's another story.
Since I don't want to install anything on the server I don't want to run a rsync server. I thought it would be enough to have the rsync executable on the notebooks and use it to synchronize all data.
This actually works quite well but it produces traffic that is not understandable for me.
Example:
I have a 25MB text file on the notebook. I sync with rsync using the command:
rsync -az --no-whole-file --progress /cygdrive/c/notebook/ /cygdrive/z/servermounted
when the file does not exist on the server the traffic produced is pretty much just 25MB.
Now, when I change a few characters in that textfile on the notebook and synchronize again it's not just synchronizing the changes but the whole file.
I get 50MB download traffic and 25MB upload traffic. Why does it produce so much download traffic? What exactly is it doing?
Can someone tell me how I get it done with less traffic, is my command wrong?
I was using the Offline Files feature of Windows XP before but that is pure crap, therefore not an option.
Thanks
Hi,
If you're running your sync on a LAN (i.e. where both computers are on the same subnet), then by default rsync will default to sending the whole file (rather than just the changes using the rsync algorithm). The rationale being that as you have the bandwidth, rsync will go for performance (at least that's my understanding of the logic). Either way ... that's what it does.
To override this behaviour, use the --no-whole-file switch
Vaughan
Hi,
To override this behaviour, use the --no-whole-file switch
Vaughan
Look at my command line above.
Hi bumba,
You're right, I read the symptoms without checking all the facts, as I got caught out with this switch myself, sorry.
I am not sure why this would be happening, possibly though because you need a client/server relationshop for cwrsync to work properly.
From a testing standpoint, I would suggest that you create a cwrsync server somewhere and run the test again. If the rsync then sends a logical amount of traffic, then you know the problem is related to trying to use a client at both ends. At that point, you'll need minds greater than mine to troubleshoot from there.
Vaughan
Hi,
Can you increase output verbosity by using -vv and publish results here ?
Rgrds Tev
This traffic is the reason that programms like cwrsync exists.Without having a server daemon running the file compare (whats changed ) need to copy the file to one CPU for comparison.
With rsync this process is divided into source and destination CPU and therefore on the network only the CRCs and Diffs are seen.
So install rsync on your server and it will work. No way ( even no other programm can solve that without causing "compare" traffic