Good day everyone,
I seem to be running into performance issues with geowebcache…possibly because I have a huge dataset, containing lots and lots of municipal structural data (a bunch of lines on a map). I seeded my cache and know that gwc works…using 900913 for my projection because I have a google maps backdrop…whenever I pan around quickly java.exe process will spike dramatically (like up to 85 % at max) – which is something I noticed before I had gwc working. I can see now that tiles are not being generated, it looks like they’re being pulled from the cache, just…not very quickly. The size of my data set shouldn’t matter, should it? Seems like it wouldn’t, since I’m just pulling up cached images.
It’s hurting a lot more when i serve up multiple layers simultaneously.
Is there anything I can do about this?
Aaron
Not sure, it should be *really* fast if you are on a fast network connection. If you have a public instance I can have a look, though there isn't all that much that can be deduced remotely.
You've seen http://www.nabble.com/gwc-re-generation-td23413870.html#a23519174 ?
If you somehow constantly trigger this race condition (things have to line up pretty well to trigger it), it could be quite slow. If so, try clearing the cache again and try the latest build, as per instructions in the post I link to above:
http://repo.opengeo.org/org/opengeo/geowebcache/1.1-SNAPSHOT/geowebcache-1.1-20090514.140726-9.jar
-Arne
Aaron_Gundel@anonymised.com wrote:
Good day everyone,
I seem to be running into performance issues with geowebcache...possibly because I have a huge dataset, containing lots and lots of municipal structural data (a bunch of lines on a map). I seeded my cache and know that gwc works...using 900913 for my projection because I have a google maps backdrop...whenever I pan around quickly java.exe process will spike dramatically (like up to 85 % at max) -- which is something I noticed before I had gwc working. I can see now that tiles are not being generated, it looks like they're being pulled from the cache, just...not very quickly. The size of my data set shouldn't matter, should it? Seems like it wouldn't, since I'm just pulling up cached images.
It's hurting a lot more when i serve up multiple layers simultaneously.
Is there anything I can do about this?
Aaron
--
Arne Kepp
OpenGeo - http://opengeo.org
Expert service straight from the developers