Each time when we get a request owl-ql loads up some Ontologies, which may have about 100 MB.
Bad idea - to load 100Mb each time. And don't forget that 100Mb data file is not the same as 100Mb data filed loaded into java objects.
I changed the heap-size to 512mb through the java option -X. Well when we geht 20 requests in a short time cocoon crashs.
20 x 100Mb ~= 2Gb. No wonder it crashes. Rewrite your app or get more RAM.
Vadim
