On 02.07.2009 08:14, J.P. Trosclair wrote:> This is just an inquiry about how rsync calculates its transfer speeds.
>
> I'm using rsync over ssh to pull data from our central office to
another
> office. Our central office is connected to the internet via T1 and the
> other office is connected to the internet via DSL. rsync is being run at
> the remote office to pull data down from the central office so I expect
> to see transfer speeds at a maximum of ~150KB/s, or so. Instead I'm
> seeing this on a ~310mb file (just one example):
>
> 0 0% 0.00kB/s 0:00:00
> 13372944 4% 12.75MB/s 0:00:23
> 22294152 7% 10.63MB/s 0:00:26
> 32173104 10% 10.23MB/s 0:00:26
>
> ....
>
> 313022664 99% 2.63MB/s 0:00:00
> 314732544 100% 6.05MB/s 0:00:49 (xfer#223, to-check=133384/557231)
>
> The transfer took 49 seconds total so there must be some truth to the
> speeds I'm seeing. I don't understand how this is being calculated
> because we don't have that kind of bandwidth on either side of the
> connection. The transfer in question is an update to the file, it
> already existed at the remote office. Is the speed a result of how rsync
> is transferring the changed file?
The transfer rate shown is the speed at which the file is syncronized,
not the rate at which the data is transfered.
If you update an existing file than rsync tries to indentify differences
and only transfers the difference, everything that doesn't need to be
transfered increases the transfer-rate.
Also if you use compression the transfer-rate refers to the speed of the
uncompressed file, so even if you transfer a new file the shown
transfer-speed can be much higher if the content has a good compression
ration.
AFAIK there is no way to see the actual transfer-rate of the data-stream
between the 2 rsync processes.
Bis denn
--
Real Programmers consider "what you see is what you get" to be just as
bad a concept in Text Editors as it is in women. No, the Real Programmer
wants a "you asked for it, you got it" text editor -- complicated,
cryptic, powerful, unforgiving, dangerous.
