[rrd-users] Incremental backup rrd file

Simon Hobson linux at thehobsons.co.uk
Tue Jul 17 10:44:27 CEST 2012

ccddtt wrote:

>1\ rsync,is a good sync file tools. but i think is not for me question.
>because: when i update some data to rrd file,the file is status is 
>'changed',so rsync need transfer the complete file .Is my analysis 

No, you don't understand how rsync works.
The main feature of rsync is it's ability to copy parts of a file. 
This makes it very efficient at updating files across a low bandwidth 
and/or high latency link (it parallelises operations to deal with 
high latency).

See http://en.wikipedia.org/wiki/Rsync and also in particular see 
external link #7 http://samba.org/~tridge/phd_thesis.pdf which is 
Andrew Tridgell's thesis on sorting from which rsync sprang.

Short version: He worked on an efficient process for sorting, and 
from that comes an efficient process for working out which bits of a 
file have changed - sorting and searching a list are somewhat related.
The basis of the rsync protocol is :
1) use the part of the process to make fast (sheap) checksums of 
blocks of the file and compare them to file blocks that are different.
2) Use the efficient search process to find if blocks have moved 
rather than been changed.
3) Copy only what's needed to re-assemble the original file at the 
destination using what's already there in the previous version.

While not applicable to your needs, rsync can even deal with a file 
where a small amount has been inserted or deleted (step 2 above). Eg 
if you add a single byte to the front of a 1GB file, then rsync will 
detect this and not copy the whole GB.

It is worth reading the man page carefully, this compression isn't 
the default for all transfer modes, nor is encryption using the SSH 
Simon Hobson

Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.

More information about the rrd-users mailing list