Jeffrey J. Kosowsky wrote:
> > But I still believe that computing pool md5sums and adding them to the
> header will add at most minimal computational overhead and will have
> many benefits. Specifically,
> 
> 1. If you use rsync with protocol 30, then you will get the md5sum for
>    free
> 
> 2. Otherwise, you only have to compute the md5sum once for each pool
>    entry. Calculating the md5sum at the time of initial pool write
>    should be fast relative to the bandwidth for transferring the file
>    in initially and relative to the disk write time to add the file to
>    the pool.
> 
> 3. Having an md5sum allows one to resolve collisions with a simple
>    compare of a 16 byte number versus reading in the entire file and
>    comparing them bitwise.
> 
> 4. Having an md5sum allows one to check pool integrity for file
>    corruption

But if you were really going to go to this much work to change to an 
incompatible format, wouldn't it be better to break files into blocks so 
all those partial matches could actually be pooled?

-- 
   Les Mikesell
    lesmikesell AT gmail DOT com


------------------------------------------------------------------------------
_______________________________________________
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/