[OSM-dev] Osmosis bug when using 'completeWays' option?

[Karl Newman]

On Dec 2, 2007 3:07 AM, Lambertus <osm at na1400.info> wrote:
> When trying to extract a number of bounding boxes from the planet file using
> Osmosis I run into a problem with the 'completeWays' option:
> Everything works well until I add the completeWays option. Osmosis fails to
> write the data to the output directory (which is on a separate harddisk) but
> instead something starts using all the space on my root disk. It is not an
> .osm file as updatedb & locate can't find any stray ones. The whole process
> runs way longer that normal and finally dies when it runs out of disk space.
> When the process dies the space on the disk is not released, that happens
> only after a reboot.
>
> Is this a bug or does the completeWays option invoke some sort of buffering?
>
> Below is an example of my Osmosis command (reduced to only one output for
> clarity), both the 'planet/' and 'extract/' directories are on a separate
> disk:
>
> java -Xmx1048m -jar utils/osmosis/osmosis.jar
> --read-xml enableDateParsing=no file="planet/planet-latest.osm"
> outPipe.0="planet"
> --tee inPipe.0="planet" outputCount=1 outPipe.0="t1"
> --bounding-box completeWays=yes inPipe.0="t1" left=3.15 bottom=50.65
> top=51.1 right=3.6 outPipe.0="bb1"
> --write-xml inPipe="bb1" file="extract/63240001.osm"
>

I wrote the completeWays patch. Yes, it does do some buffering--it has
to keep all the nodes in order to make the ways (which come later)
complete. It writes it to a compressed (gz) temp file. I'm sorry, I
never tested it on a complete planet file. If you're working on a
particular area area, you might want to start with a simple bounding
box first (no completeWays) to limit the data that is buffered.

By the way, one of my future projects is to create a "tiling" task for
Osmosis which will do exactly what you are wanting.

Karl