I have a lengthy osm2pgsql import that failed due to insufficient memory. I have added 16GB af cache and would like to resume the process if possible. Based on the following, is this possible or do I need to restart? osm2pgsql -d gis --create --slim -G --hstore --tag-transform-script ~/src/openstreetmap-carto/openstreetmap-carto.lua -C 2000 --number-processes 1 -S ~/src/openstreetmap-carto/openstreetmap-carto.style ~/data/europe-latest.osm.pbf asked 16 Jul '18, 14:09 Jacknj |
I would definitely switch to --flat-nodes if you run out of memory, as this will store the nodes on disk instead of RAM. Especially with SSD, you shouldn't worry about slow imports with --flat-nodes. I recently succeeded to load the whole of Europe on a 12GB RAM virtual machine with plenty of disk space, but only after I dropped inclusion of advanced settings like cache, and simply let osm2pgsql itself figure out its optimal processing setting. It took just some 17 hours on a Core i7 laptop. See here for a few more details and observations: answered 16 Jul '18, 17:48 mboeringa |
Not easily, no.
permanent link
This answer is marked "community wiki".
answered 16 Jul '18, 15:51 SomeoneElse ♦ |