From: <Alb...@hz...> - 2008-04-30 07:26:41
|
Yes, i made the JVM settings in the tomcat (max heap space 1024MB - i think tomcat doesn't manage more than this). I know that it's hard to debug an application remotely ;o). I had to do so too with my ArcMap2SLD-app sometimes. For me these tests were successful, because finally i was able to connect to an ArcSDE backend. I think we will apply geoserver here. Now we know about the restrictions concerning SDE backend and we will align the use of geoserver with the given restrictions (which are o.k.). We are establishing that ArcSDE 9.2 instance with all data what was formerly running on 8.3 instance. If the problem will appear again, i will get back and inform you about it. If not let's say it was a problem of different versions. Best wishes Albrecht > Hi Albercht, sorry for the late response > > also less than 4G should be enough, provided you're saying to > the JVM to use > them. But I still need to sit down and seriously look at this > issue, which I > didn't have time to yet. > > sorry I know its frustrating, but I still don't know if its a > bug in the > arcsde plugin, a bug in the esri jars, just the nature of the > beast of what > else is going on. > All I can say is the only way we can get the list of layers > from sde is by > asking an sde connection for the list of layers, which returns a > java.util.Vector(SeLayer) and I would be surprised a list of > 300 SeLayer > objects need to take that much memory, so hopefully there's > something else > wrong. > > In the meantime I can only think if you would be able to test > connecting to an > instance with those 300+ feature types from another machine, > and ideally a > different sde instance, may be its just some weird environment stuff ? > > regards, > > Gabriel |