[OSM-dev] Evaluating the Full-History-Dump

Peter Körner osm-lists at mazdermind.de
Thu Oct 14 21:14:38 BST 2010


Am 12.10.2010 00:52, schrieb Anthony:
 > On Mon, Oct 11, 2010 at 5:07 PM, Peter 
Körner<osm-lists at mazdermind.de>  wrote:
 >> These dumps could be used as test-cases to benchmark tools and
 >> compare results without the need of a huge 64GB RAM server or 100
 >> MBit Internet connection.
 >
 > FWIW, I've been able to download and parse the full history dumps on
 > my less than a gig of ram server connected to a cable modem.  It took
 > about a week and half IIRC.

My proplem is not processing the full dump with a ready tool but testing 
a tool under development with the dump. It always take minutes to skip 
those >5000k changesets.

I wrote a tool [1] that splits a planet or full-experimental dump into 4 
files:
  - nodes.osm.bz2
  - pois.osm.bz2
  - nodes+ways.osm.bz2
  - nodes+ways+relations.osm.bz2

It best runs on a multicore system (1 de-, 4 compress and 1 splitter). I 
currently run it on a full-experimental dump and I think it will take 
some more hours but it would be useful to share them on planet.osm.org.

Peter


[1] <http://svn.toolserver.org/svnroot/mazder/split-osm/>



More information about the dev mailing list