On Fri, Feb 8, 2013 at 2:24 AM, Sorin Srbu <sorin.srbu AT orgfarm.uu DOT se> 
wrote:
>>
>> > This indicates that I was wrong; directory/inode traversal is not the
>> > issue. It's speedy enough to run the test. It is likely in the rsync
> block
>> > checksum comparison after all.
>>
>> How does this time compare to an incremental where few/no files
>> actually change?  You should only do the checksum comparison on fulls
>> or files where the directory timestamp/length differ on incrementals.
>>   If there is an extreme difference, maybe you are out of available
>> RAM and pushing the server or target into swap with the size of the
>> directory contents.
>
> "A lot faster".

During the 'a lot slower' incremental, does 'top' show swap being used
on either end?    Do the backuppc statistics show that in fact no
files were transferred?

> So can I disable checksumming and still use rsync, or is it tar I want?

Checksumming is kind of the point of rsync, but in the case of
incrementals is should only happen on files where the timestamp or
length differ.  So it isn't causing the problem of slow incrementals
if there are no files changed since the last full run (and note that
the comparison is against the last full, unless you are using
incremental levels which will increase the load on the server).

-- 
   Les Mikesell
       lesmikesell AT gmail DOT com

------------------------------------------------------------------------------
Free Next-Gen Firewall Hardware Offer
Buy your Sophos next-gen firewall before the end March 2013 
and get the hardware for free! Learn more.
http://p.sf.net/sfu/sophos-d2d-feb
_______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List:    https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:    http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/