[OSM-dev] OpenStreetMap Carto release v3.1.0
Dear all, Today, v3.1.0 of the openstreetmap-carto stylesheet (the default stylesheet on openstreetmap.org) has been released. Changes include - Added coffee shop rendering - Added health clinic rendering - Adjusted place label typography - Road shield rendering improvements - Internal code cleanups Thanks to all the contributors for this release. For a full list of commits, see https://github.com/gravitystorm/openstreetmap-carto/compare/v3.0.1...v3.1.0 As always, we welcome any bug reports at https://github.com/gravitystorm/openstreetmap-carto/issues. ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] size of postgresql database - without flat file
On 1/28/2017 3:00 PM, Walter Nordmann wrote: i'm just installing my brand new server to get my applications (missing boundaries and many more) running again. The server has 2x 960GB SSD and 1x 2TB SATA disk installed. I'm doing a full planet import using osm2pgsql 0.92. This time i'm doing the import without flat file, because i would like to use the nodes in planet_osm_nodes for some new queries. But: planet_osm_nodes got an actual size of 785 GB, which is much much more than i exspected. Is that the usual size or may be, i'm doing something wrong? With the version of osm2pgsql you have when --extra-attributes is specified then tags are added to each object with the username, uid, version, timestamp, and changeset. This adds about 100 bytes per node to the table, which should add about 350GB. I would expect the size of planet_osm_nodes with a plain --slim import to be about 175GB, plus indexes. Does your 785GB figure include indexes? If so, this sounds about right. The latest versions of osm2pgsql do not have this issue because the node slim table schema has changed to no longer store tags, because they are never needed. It's worth noting that 100 bytes per object for metadata is remarkably inefficient because it gets stored in a text[] array like "osm_user,lokejul,osm_uid,2034065,osm_version,11,osm_timestamp,2015-04-12T18:39:22Z,osm_changeset,30169891". This uses more bytes because it stores "tag" names, and because there are more efficient ways to store numbers and timestamps than as text. ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev
[OSM-dev] size of postgresql database - without flat file
Hi, i'm just installing my brand new server to get my applications (missing boundaries and many more) running again. The server has 2x 960GB SSD and 1x 2TB SATA disk installed. I'm doing a full planet import using osm2pgsql 0.92. This time i'm doing the import without flat file, because i would like to use the nodes in planet_osm_nodes for some new queries. But: planet_osm_nodes got an actual size of 785 GB, which is much much more than i exspected. Is that the usual size or may be, i'm doing something wrong? my import batch: - # no more Flat File # + extra_attributes set -x cd /osm/db/$1/create OSM2PGSQL=/usr/local/bin/osm2pgsql $OSM2PGSQL --verbose \ --create \ --slim \ --exclude-invalid-polygon \ --style /osm/db/wno_2017.style \ --port 5432 \ --database $1 \ --latlon \ --username postgres \ --hstore-all \ --hstore-add-index \ --tablespace-main-data $1_ttmp \ --tablespace-main-index $1_is1 \ --tablespace-slim-data $1_ttmp \ --tablespace-slim-index $1_is2 \ -C 25000 \ --cache-strategy optimized \ --extra-attributes \ --number-processes 12 \ --keep-coastlines \ --multi-geometry \ import/$2 -- called: do_import planet3 planet-latest.osm.bz2 any hints? Regards walter ___ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev