Up to now I've been downloading the UK and Ireland extracts of the OSM
planet and feeding them to osm2pgsql in "memory" mode. Unfortunately, the
combination of the two is now too much for osm2pgsql to hold in memory (the
virtual memory requirement exceeds the maximum per-process 3GB allocation of
a 32 bit Linux environment, even though its actual use of memory is only
1.5GB). The UK section on its own looks like it will grow to exceed that
within a month or two.
For the moment, I can separately run "osm2pgsql -c ..." for one section and
"osm2pgsql -a ..." for the other, and still keep things fitting in memory.
I can guess what you'll say: use osm2pgsql in slim mode. Yes, that's a
solution for a single section, but for two sections with a little overlap
you hit a problem with the database throwing an error due to duplicate
primary keys.
Has anyone worked out how to get round this problem? Would osmosis be
useful instead?
--
Paul Martin <
pm@...>
_______________________________________________
GPSdrive mailing list
GPSdrive@...
http://lists.gpsdrivers.org/mailman/listinfo/gpsdrive
