Hi Torben, Have a look at how much memory each individual task is requiring as it runs, it's memory use will grow through to when it builds the last level into the quad tree/pyramid that it's handling. It could be that one task takes up lots of memory, but runs fine, but run multiple osgdem in a parallel on one machine and it could blow the memory.
To cut the memory load down you can change the split levels that it's building against, or just change your machines file to tell vpbmaster to run less osgdem processes in parallel on each machine. Another problem that can occur is if your file system isn't set up to handle asynchronous writes to it and gets out of sync. Robert. On Mon, Oct 5, 2009 at 5:39 PM, Torben Dannhauer <[email protected]> wrote: > Hi Robert, > > executing the taskfiles manually works without crash. > > If I would write a script which analyses all taskfiles regarding their > pending status, and execute that taskfile if it's status is "pending" - would > it result in a usable database? > > Because osgdem itselfs works properly I assume that the high system load of > vpbmaster and all of that open files are the reason for the crash. Especially > that several independet libraries crash intidates a problem in my > configuration/systemload. > For example I rised the ulimit of allowed open files to 500000, but the > problem of vpbmaster remains. > > Are there other limits/values (listable via ulimit -a) that could cause the > problem? > > > Thank you very much! > > Best regards, > Torben > > ------------------ > Read this topic online here: > http://forum.openscenegraph.org/viewtopic.php?p=17905#17905 > > > > > > _______________________________________________ > osg-users mailing list > [email protected] > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > _______________________________________________ osg-users mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
