[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: ftpsync README rewrite

Joerg Jaspert (joerg@debian.org) wrote on 17 September 2009 10:20:
 >On 11874 March 1977, Lee Winter wrote:
 >> The other thing I accomplished is to find a way to eliminate the load
 >> that rsync imposes on upstream mirrors.  IMO it would not take much
 >> work to tweak rsync and adjust the calling scripts so that the
 >> upstream mirrors could be completely passive.

The problem with this approach is that it needs changes (non-trivial
ones, btw) to rsync. That doesn't "catch" because mirrors change their
versions very slowly. An example is that version 3 is much better but
still not widely used.

 >If you find a way that actually works nicely I am MOST interested to
 >learn it!

Thanks to your improvements to the trace file:

% cat project/trace/ftp.br.debian.org 
Thu Sep 17 22:42:06 UTC 2009
Used rsmir version 1
Running on host: ftp.br.debian.org

The sender must still calculate checksums since it uses rsync (the rs
in rsmir) but this is neglible nowadays. The (by far) worst hit is the
scanning of the entire distribution both on the client and server. And
with ftpsync it's done twice! Oh no...

rsmir doesn't do the scanning on either end, while still having
the mandatory features of a debian mirror.

Most mirrors don't have problems, so they won't do any signficant
change. Only the main mirrors and the master are hard pressed; they
might change but without the others it won't help much. So I'm rather
firmly convinced that significant improvements are extremely unlikely.
Things will end up improving only by the increase in the RAM amount of
the machines.

BTW, the quality of the Debian repository and its management is
impressive. It's the best I've seen. It's the only one that allows
what rsmir does without changes. Even other distros that have copied
the structure managed to break it...

Reply to: