This is a static archive of our old OSM Help Site. Please post any new questions and answers at community.openstreetmap.org.

Tile cache update frequency

0

Hi, I am having a customized OSM setup with our own database. Currently I'm getting the data as per my requirement edited by iD editor. Now I want this data to be rendered as map. Also I want to cache the generated tiles of this data. So some requirement:

  1. I want to know standard practices for generating map using above data. Like taking .osm data offline, do some processing, make a shapefile, publish it in GeoServer or is there any other way also?

  2. Since tile cache is a suitable option for faster data/tile rendering, what are the practices to generate tiles and how to update these tiles as data will be modified/added/updated regularly through iD editor.

Thanks in advance.

asked 09 Jan '14, 11:33

GoldenCompass's gravatar image

GoldenCompass
40●3●3●6
accept rate: 0%

edited 09 Jan '14, 14:59

iii's gravatar image

iii
4.9k●8●40●82

(just in case it isn't clear) from previous questions I believe that the questioner has their own copy of the rails port deployed.

(09 Jan '14, 11:39) SomeoneElse ♦

@SomeoneElse Yes, I am having a copy of OSM setup (as per my requirement) up and running and getting data from concerned users.

(09 Jan '14, 11:44) GoldenCompass

One Answer:

0

IMHO you need to decide which way you want to use to create your tiles:

  1. Rendering on demand
    So your TMS get's the client tile requests and renders/splits the desired area on the fly.
  2. Preprocessing all tiles
    All tiles are rendered and then only the affected areas (by your edits) get updated and rerendered. This takes a lot of space (if you need to provide it as global service), but it scales better if you have a larger amount of clients.

Both ways can be combined with mapcaches, but of course this limits the tile update frequency as the way to the client get's longer and the caches need time to propagate changes and to get flushed.

In every case you need to create a PostGIS DB that suits the rendering stack requirements (AFAIK this is a different one from the central rails port). Then you setup the Mapnik renderer (or a different one) and decide if you call them after edits or if you get a client request. A good tutorial is www.switch2osm.org also you might read http://wiki.openstreetmap.org/wiki/Creating_your_own_tiles

answered 09 Jan '14, 15:10

iii's gravatar image

iii
4.9k●8●40●82
accept rate: 10%