[
http://issues.apache.org/jira/browse/DERBY-1142?page=comments#action_12420461 ]
Daniel John Debrunner commented on DERBY-1142:
----------------------------------------------
Two patches applied (420656 & 421015) that stop the memory leak shown in the
repro.
A leak still exists if the repro is modified to not close the ResultSet.
This is because the activations marked unused in the language connection
context are only cleaned up if a commit or rollback is executed.
In this code no such call is made.
One solution is to look for unused activations when an activation is added to
GenericLanguageConnectionContext (addActivation method).
A couple of possible solutions to avoid overhead when no check is needed are:
- Have Activation.markUnused() set an boolean (unsynchronized) flag
in the LCC to indicate it has some unused activations
- Only perform the check if the number of activations is beyond
some threshold.
Not sure how urgent it is needed to solve this issue, if the application calls
commit or rollback in any way (as most applications will at some point)
then memory leak will not be seen. This repro is an extereme edge case.
> Metadata calls leak memory
> --------------------------
>
> Key: DERBY-1142
> URL: http://issues.apache.org/jira/browse/DERBY-1142
> Project: Derby
> Type: Bug
> Components: JDBC
> Versions: 10.1.2.1, 10.2.0.0
> Reporter: Knut Anders Hatlen
> Assignee: Daniel John Debrunner
> Priority: Minor
> Attachments: 1142_close_single_use_activations_draft.txt, metadataloop.java
>
> When calling a DatabaseMetaData method that returns a ResultSet,
> memory is leaked. A loop like this (using the embedded driver)
> while (true) {
> ResultSet rs = dmd.getSchemas();
> rs.close();
> }
> will eventually cause an OutOfMemoryError.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
