Kristian Waagan wrote:
John Embretsen wrote:
Hi Ture,
[ snip ]
I didn't think it was this easy to fill up the heap (by not explicitly
closing Statement objects) anymore (see e.g. DERBY-210), but there are
obviously some vulnerabilities left. So thank you for sharing your
code and for reporting this!
Hi John,
I think the reason why the OOME happened so fast in this application was
because of the sorting (ORDER BY). I don't know enough about how sorting
is implemented in Derby to have an opinion, but I think we should have a
look at it.
I will do a little pre-investigation and create a Jira if appropriate.
OK, thanks! From quickly reading the thread I was under the impression that
adding a call to Statement.close() solved the problem, but I guess that was only
part of it.
One of the DOTS tests I've been running also stumbled across a memory leak
related to sorting, see http://issues.apache.org/jira/browse/DERBY-2176 .
Interesting coincidence. That was fixed with 10.2.2.0, though, so this must be
something else.
--
John
