[mapserver-users] The Case of the Disappearing Roads

Rahkonen Jukka jukka.rahkonen at mmmtike.fi
Tue Aug 13 00:00:36 PDT 2013


Hi,

Have you tried to use ogrtindex http://www.gdal.org/ogrtindex.html instead of tile4ms? Try also if there is any difference if you read the tileindex through ogr or with the native Mapserver method. I have found some differencies with these options. For example I found this note from my mapfiles (but Even Rouault could not reproduce my SDO_BODY error).
#CONNECTIONTYPE OGR #with this SDO_BODY does not work on Linux build (CentOS)
  TILEINDEX "/ms4w/apps/ogrtindextest.shp,0"

-Jukka Rahkonen-

Joseph Marlin wrote:

> I am indeed using tile indices. I didn't think there would be any problem with that
> step, but just to be sure I went ahead and checked. They all looked ok, but I
> noticed that all the features that use tile indexes were the exact features that
> aren't showing up- small roadways, highway shields, etc. This was highly
> suspicious, and so I did a lot of digging.
> 
> I recreated all the index shape files, this time using an old version of the tile4ms
> and shptree utilities that I had fortunately saved when I generated the map in
> 2011, and found that tileindex shapefiles were no much larger, and the map was
> rendered correctly!
> 
> So, all I can say for sure is that the current version of tile4ms and shptree
> definitely did not work. This issue has been reported at least twice before it
> seems,
> http://trac.osgeo.org/mapserver/ticket/4259
> http://www.mail-archive.com/mapserver-
> users at lists.osgeo.org/msg17207.html
> 
> And I guess it is worth noting that this problem is still present in the latest
> version (MapServer version 6.2.1 OUTPUT=PNG OUTPUT=JPEG SUPPORTS=AGG
> SUPPORTS=FREETYPE SUPPORTS=ICONV INPUT=JPEG INPUT=SHAPEFILE).
> 
> Thanks so so much for your help! Let me know if you need further info.
> Joseph
> 
> 
> 
> ----- Original Message -----
> From: "thomas bonfort" <thomas.bonfort at gmail.com>
> To: "Joseph Marlin" <jmarlin at saucontech.com>
> Cc: "MapserverList OSGEO" <mapserver-users at lists.osgeo.org>
> Sent: Saturday, August 10, 2013 3:18:03 AM
> Subject: Re: [mapserver-users] The Case of the Disappearing Roads
> 
> Another thing to check: are you using a tileindex to reference your shapefiles,
> and if so, is the tileindex up-to-date (i.e. does it reference all your shapefiles).
> 
> --
> thomas
> 
> On 9 August 2013 20:04, Joseph Marlin <jmarlin at saucontech.com> wrote:
> > Thanks Thomas for the idea. I think I understand what you're asking. We pre-
> render all our tiles though, so while the screenshots are indeed showing a
> javascript viewer, that viewer is simply loading the prerendered images,
> retrieved from tilecache. And I have made sure that it isn't TileCache storing old
> tiles.
> >
> > ----- Original Message -----
> > From: "thomas bonfort" <thomas.bonfort at gmail.com>
> > To: "Joseph Marlin" <jmarlin at saucontech.com>
> > Cc: "MapserverList OSGEO" <mapserver-users at lists.osgeo.org>
> > Sent: Friday, August 9, 2013 1:54:41 PM
> > Subject: Re: [mapserver-users] The Case of the Disappearing Roads
> >
> > Joseph,
> > are the images you are posting a direct result of a getmap request
> > (i.e. with width=1007&height=454), or are they a screenshot of a
> > javascript client that is doing tiled requests to mapserver?
> > If those are tiled requests, are your layer/class minscale/maxscale
> > settings set to the exact values of your requested resolutions (i.e.
> > their might be some rounding errors in that case that make a
> > scale-dependant class appear or disappear based on floating point
> > rounding errors).
> >
> > --
> > thomas
> >
> > On 9 August 2013 19:03, Joseph Marlin <jmarlin at saucontech.com> wrote:
> >> The Case of the Disappearing Roads
> >>
> >> We create shapefiles with SQL querying world data that's been loaded into a
> PostgreSQL database. A python script requests tiles from mapserver at different
> zoom levels, and we generate the entire map like that.
> >>
> >> Now, at the sixth zoom level, level 3 roads (medium size roads) should be
> rendered. However, they are actually rendered only in a small section of the
> world, a box bounded by Toronto in the northwest, Cleveland in the southwest,
> and the Atlantic in the east. Elsewhere, level 3 roads are not rendered at all.
> >>
> >> As you can see in this image: http://i.imgur.com/6McvGOJ.png, the small
> gray roads, level 3 roads, (and highway shields, for that matter) that are visible
> on the top half suddenly stop being rendered by mapserver.
> >>
> >> In all the following more zoomed in levels, no roadways smaller than
> >> level 2 are rendered at all anywhere, as you can see here:
> >> http://i.imgur.com/8MTcPoi.png. I've verified that the shapefiles
> >> contain the data on the smaller roads as I have viewed them just fine
> >> with QGIS, as you can see here: http://i.imgur.com/S3W3Iy8.png
> >>
> >> I'm so confused as to what could possibly make roads disappear midway
> through a level. If it was a problem with my mapfile, why would they show up in
> part of Northeast US, but not anywhere else? Where do I even start looking to
> solve this?
> >>
> >> Thanks so much!
> >>
> >> Additional info:
> >> SRS: EPSG:900913
> >> Zoom levels: 19567.8792375, 9783.93961875, 4891.969809375,
> >> 2445.9849046875, 1222.99245234375, 611.496226171875,
> >> 305.7481130859375, 152.87405654296876, 76.43702827148438,
> >> 38.21851413574219, 19.109257067871095, 9.554628533935547,
> >> 4.777314266967774, 2.388657133483887,  1.1943285667419434,
> >> 0.5971642833709717, 0.29858214168548586 Bounding box:
> >> -20037508.3427892,-20037508.3427892,
> >> 20037508.3427892,20037508.3427892
> >> _______________________________________________
> >> mapserver-users mailing list
> >> mapserver-users at lists.osgeo.org
> >> http://lists.osgeo.org/mailman/listinfo/mapserver-users
> _______________________________________________
> mapserver-users mailing list
> mapserver-users at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-users


More information about the mapserver-users mailing list