[Tilecache] Tilecache too slow

[Nico Mandery]

Hi Stefan,

maybe this is of some help:
http://trac.mapnik.org/wiki/OptimizeRenderingWithPostGIS

regards,

nico

On 31.05.2010 17:25, Stefan Priess wrote:
>
> Hello List,
>
> in loaded the planet.osm file into my db and want to prerender some 
> tiles so i call:
>
> tilecache_seed.py osm 0 7
>
> my cifg:
>
> [osm]
> type=Mapnik
> mapfile=/opt/mapniktools/osm.xml
> extension=png
> spherical_mercator=true
> metabuffer=256
> metatile=yes
>
> Output:
>
> Cache miss: -20037508.34,-20037508.34,20037508.3384,20037508.3384, 
> Tile: x: 0, y: 0, z: 0, time: 4.50857400894
> Cache miss: 
> -20037508.34,-20037508.34,-0.000799998641014,-0.000799998641014, Tile: 
> x: 0, y: 0, z: 1, time: 2.24050211906
> Cache miss: -20037508.34,-20037508.34,-10018754.1704,-10018754.1704, 
> Tile: x: 0, y: 0, z: 2, time: 47.7445428371
> Cache miss: -20037508.34,-20037508.34,-15028131.2552,-15028131.2552, 
> Tile: x: 0, y: 0, z: 3, time: 33.1307868958
> Cache miss: 5009377.084,-20037508.34,10018754.1688,-15028131.2552, 
> Tile: x: 5, y: 0, z: 3, time: 23.2806358337
> Cache miss: -20037508.34,5009377.084,-15028131.2552,10018754.1688, 
> Tile: x: 0, y: 5, z: 3, time: 27.7289340496
> Cache miss: 5009377.084,5009377.084,10018754.1688,10018754.1688, Tile: 
> x: 5, y: 5, z: 3, time: 21.4363980293
> Cache miss: -20037508.34,-20037508.34,-17532819.7976,-17532819.7976, 
> Tile: x: 0, y: 0, z: 4, time: 2.18496513367
> Cache miss: -7514065.628,-20037508.34,-5009377.0856,-17532819.7976, 
> Tile: x: 5, y: 0, z: 4, time: 1.23492002487
> Cache miss: 5009377.084,-20037508.34,7514065.6264,-17532819.7976, 
> Tile: x: 10, y: 0, z: 4, time: 0.98893404007
> Cache miss: 17532819.796,-20037508.34,20037508.3384,-17532819.7976, 
> Tile: x: 15, y: 0, z: 4, time: 0.568192005157
> Cache miss: -20037508.34,-7514065.628,-17532819.7976,-5009377.0856, 
> Tile: x: 0, y: 5, z: 4, time: 23.5294082165
>
> This is too slow! 8GB Mem and 2x CPU with 4x Cores each 2,66GHz should 
> have enough power.
> I use Tilecache 2.10 with the Patch to avoid the zero data layer bug.
>
> Any ideas?
>
> regards
>
> Stefan 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/tilecache/attachments/20100602/eccf7891/attachment.html