I have multiple instances of my application for my users. Each user can specify an area which I want to load in my own PostgreSQL server, that make it's possible to render maps with my own mapnik stylesheet and nik4.
The idea was to load the data via overpass-api.de and load it with osm2psql. Each region has a size about 150MB. To separate the data, I want to use the --prefix option. Then I thought I could download the region each month again, use --append instead of --create option in osm2pql and the data will be updated. But as I know now, that is not possible.
A possible solution is, that I drop the tables and load it again. Another could be osmosis. If I don't delete the downloaded file I could create a changeset that is possible to load with the append option. But I have to save the old download file. There is a write write-pgsql-change option but it seems not to work with this multiple instances of the osm2psql import.
Is there any other solution to solve this problem or maybe I'm on a complete wrong way - so I'm open for any suggestions.
PS: I'm not able to load the whole planet, due to hardware limitations. At the moment, there are only 3 users, so the maintaining is quite small at the moment. But in future the will be a bit more users, and I don't know which region they choose, so I'm not able to download some geofabrik extracts and update them daily. And it seems much faster to handle much smaller tables with smaller extracts.