Hello people,

I've been doing some tests on the 1.5.1 version regarding the extreme
amount of memory that cloud.getList() seems to allocate. It turns out
that if I issue a multilevel query that returns 10 Nodes, and repeat
this 10,000 times, the amount of memory the JVM needs grows with about
10 Mb and doesn't go down again, crashing the JVM after about 60,000
queries depending on configuration.

This only happens when the multilevel cache is on, otherwise memory
usage stays even. Which turns out not to be much slower anyways - at
least for very simple queries.

Attached is a JSP page that demonstrates the problem.

Does anyone have a solution / patch for this problem? We do need the
multi-level queries, so our only option now is to disable the
multi-level cache.

Greets,

Joost Diepenmaat
Framfab Nederland BV







-- 


Reply via email to