MapBox

MapBox maintains a Public Data Set with the complete collection of OpenStreetMap data stored as a PostgreSQL data cluster. You can get access to an enormous amount of geographic data in a matter of seconds by creating an EBS volume based on this snapshot. Traditionally The Planet was bzip2 compressed, so it unzips from 7.4GB to 134GB, and then occupies ~80GB in a PostGIS database.

Preparing the data set

If you haven't done so already, start a MapBox EC2 instance.

Make a note of the availability zone and the ID of the instance. You'll need them when you create and attach the EBS volume.

Create a new EBS volume based on the OSM snapshot.

Attach the volume to your EC2 instance. You may use any valid device name in place of dev/sdx.

Connect to your EC2 instance using SSH. Once you're logged in create a mountpoint for your volume. I've selected /mnt/osm as my mountpoint, but you can call this whatever you like. Once you have a mountpoint, go ahead and mount your volume.

Your mount point should now contain the following directories:

The volume we just created contains a PostgreSQL data directory, and we'll need to create a new data cluster using this directory.

The output from this command will contain the port number that was assigned to the data cluster. Make a note of it because you'll need to specify the port number to connect to the server.

Now, restart the PostgreSQL server to start using the data.

Using the OSM Planet EBS with other EC2s

To use the OSM Planet EBS with AMIs from other sources is simple - the only requirement is PostgreSQL 8.3 or newer, which can be installed with a package manager on Linux or an installer package on Windows. The EBS contains the PostGIS extensions within the database dump, so it is not necessary to install PostGIS.