Deepa Remesh wrote:
Thanks for the update. Here is how I interpreted the test results from
John's mails and comments at
https://issues.apache.org/jira/browse/DERBY-210#action_12364800
http://www.nabble.com/Re%3A-jira-Updated%3A-%28DERBY-210%29-Network-Server-will-leak-prepared-statements-if-not-explicitly-closed-by-the-user-until-the-connection-is-closed-p2783155.html
http://www.nabble.com/Re%3A-jira-Updated%3A-%28DERBY-210%29-Network-Server-will-leak-prepared-statements-if-not-explicitly-closed-by-the-user-until-the-connection-is-closed-p3279845.html
* Without any patches, DOTS test gives first OOME after 3 hours.
* With my first draft patch, DOTS test ran upto 55 hours before giving an OOME.
* With patch5, DOTS test ran upto 56 hours before giving an OOME.
If the above statements are correct, I think patch5 is improving the
memory usage. It solves the specific problem of memory leak of
prepared statements. It looks like Derby client driver still has some
more areas which can be possibly improved. If this is the case, I
think it would be good to get patch5 committed and we can open new
JIRA issues to track the remaining OOMEs.
Your interpretation is correct, as far as I can see. I join you in
recommending that patch5 for DERBY-210 gets committed, based on the test
results I've seen so far. Your suggestion of opening new Jira issues for
remaining issues sounds good to me.
I will nevertheless continue to run more tests, especially since one of my
JVM settings seems to have affected the outcome of my first tests of patch5,
as previously mentioned in this thread.
--
John
