Oracle AQAPI causing memory leakage

Question

Good day
I am having a memory leakage on my production server which degrades and eventually dies, the cause is a job that is executed every five minutes which polls the Oracle Queue using the AQAPI. I can see that garbage collector is attempting to finalize() the AQOracleQueue, but this object is waiting on a lock.
My code is as follows:
package com.strysie.aq;
import java.sql.Connection;
import oracle.AQ.AQDequeueOption;
import oracle.AQ.AQDriverManager;
import oracle.AQ.AQMessage;
import oracle.AQ.AQQueue;
import oracle.AQ.AQSession;
*public class QueueDAO {*
*public QueuePayload dequeueMsg() throws Exception {*
AQSession session = null;
AQQueue queue = null;
Connection connection = null;
*try {*
connection = connectionFromPool();
connection.setAutoCommit(false);
session = AQDriverManager.createAQSession(connection);
queue = session.getQueue("mySchemaName", "myQueueName");
AQDequeueOption dequeueOption = new AQDequeueOption();
dequeueOption.setConsumerName("myConsumerName");
String browseMode = System.getProperty("AQBrowseOrDequeue");
*if (browseMode.compareToIgnoreCase("Dequeue") != 0) {*
dequeueOption.setDequeueMode(AQDequeueOption.DEQUEUE_BROWSE);
AQMessage message = queue.dequeue(dequeueOption, new QueuePayload());
return (QueuePayload) message.getObjectPayload().getPayloadData();
*} catch (Exception ex) {*
throw ex;
*} finally {*
*try {*
*if (queue != null) {*
queue.close();
*} catch (Exception e) {*
e.printStackTrace();
*try {*
*if (session != null) {*
session.close();
*} catch (Exception e) {*
e.printStackTrace();
*try {*
*if (connection != null) {*
connection.commit();
*} catch (Exception e) {*
e.printStackTrace();
returnToPool(connection);
*private Connection connectionFromPool() {*
Connection connection = null;
*// fetch connection from pool*
return connection;
*private void returnToPool(Connection connection) {*
*// return connection to pool*
package com.strysie.aq;
import java.math.BigDecimal;
import java.sql.SQLException;
import oracle.jdbc.driver.OracleConnection;
import oracle.sql.CustomDatum;
import oracle.sql.CustomDatumFactory;
import oracle.sql.Datum;
import oracle.sql.STRUCT;
import oracle.sql.StructDescriptor;
*public class QueuePayload implements CustomDatum, CustomDatumFactory {*
public String account;
public BigDecimal amount;
static final QueuePayload message_t_Factory = new QueuePayload(null, null);
*public static CustomDatumFactory getFactory() {*
return message_t_Factory;
*public QueuePayload() {*
*public QueuePayload(String account_number, BigDecimal amount) {*
this.account = account_number;
this.amount = amount;
*public Datum toDatum(OracleConnection c) throws SQLException {*
StructDescriptor sd = StructDescriptor.createDescriptor("QUEUEPAYLOAD", c);
*Object[] attributes = { account, amount};*
return new STRUCT(sd, c, attributes);
*public CustomDatum create(Datum d, int sqlType) throws SQLException {*
*if (d == null) {*
return null;
Object[] attributes = ((STRUCT) d).getAttributes();
return new QueuePayload((String) attributes[0], (BigDecimal) attributes[1]);
** @return the account_number*
*public String getAccount() {*
return this.account;
** @return the amount*
*public BigDecimal getAmount() {*
return this.amount;
** @param account_number*
** the account_number to set*
*public void setAccount(String account_number) {*
this.account = account_number;
** @param amount*
** the amount to set*
*public void setAmount(BigDecimal amount) {*
this.amount = amount;
When I extract a thread dump I encounter the following on the Finalizer thread:
*"Finalizer" daemon prio=10 tid=0x00002aaab31c7800 nid=0x7469 waiting for monitor entry [0x0000000040a32000]*
java.lang.Thread.State: BLOCKED (on object monitor)
at oracle.jdbc.driver.OracleStatement.close(OracleStatement.java:1341)
- waiting to lock <0x00000007b0745288> (a oracle.jdbc.driver.T4CConnection)
at oracle.AQ.AQOracleQueue.close(AQOracleQueue.java:3225)
at oracle.AQ.AQOracleQueue.finalize(AQOracleQueue.java:3364)
at java.lang.ref.Finalizer.invokeFinalizeMethod(Native Method)
at java.lang.ref.Finalizer.runFinalizer(Finalizer.java:83)
at java.lang.ref.Finalizer.access$100(Finalizer.java:14)
at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:160)
Please assist.
Best regards
Rudi

Answer 1

Can only 'guess' since the code is missing that shows how the connection and statement instances are being created, pooled and used.
Your code is not setting the 'wait' option for the dequeue and the default is DEQUEUE_WAIT_FOREVER for a message. So my 'guess' is that an instance of the code you show is 'waiting' for a message and keeping it's connection and statement open and then another thread is trying to use that connection and it is locked or another thread is trying to shutdown the application when the problem appears.
See Dequeuing Messages in the JDBC Developer's Guide
http://docs.oracle.com/cd/B28359_01/java.111/b31224/streamsaq.htm#BABJGBBA
>
•Wait: Specifies the wait time for the dequeue operation, if none of the messages matches the search criteria. The default value is DEQUEUE_WAIT_FOREVER indicating that the operation waits forever. If set to DEQUEUE_NO_WAIT, then the operation does not wait. If a number is specified, then the dequeue operation waits for the specified number of seconds.
>
Try setting the 'wait' option to DEQUEUE_NO_WAIT or to a number of seconds and see if the problem reoccurs.
NOTE: If this problem is occuring because another thread is trying to use this same connection when it is locked it probably means you have a problem in your connection pooling code since it would mean the connection pool is allowing another thread to use a connection that is already being used.