[ 
https://issues.apache.org/jira/browse/DBCP-484?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16410711#comment-16410711
 ] 

Zheng Feng commented on DBCP-484:
---------------------------------

Hi,

I think there might be a race condition when closing the ManagedConnection

There are two methods which could call connection.close() and the one is in 
[ManagedConnection.close()|https://github.com/apache/commons-dbcp/blob/master/src/main/java/org/apache/commons/dbcp2/managed/ManagedConnection.java#L160]
 and the other is 
[ManagedConncetion.transactionComplete()|https://github.com/apache/commons-dbcp/blob/master/src/main/java/org/apache/commons/dbcp2/managed/ManagedConnection.java#L188]
 which is used in calling back when the transaction is completed.

{code}
public void close() throws SQLException {
        if (!isClosedInternal()) {
            try {
                // Don't actually close the connection if in a transaction. The
                // connection will be closed by the transactionComplete method.
                if (transactionContext == null) {
                    super.close();
                }
            } finally {
                setClosedInternal(true);
            }
        }
    }
{code}

The thread of close() could be switched after check (transactionContext == 
null) and before the setCloseInternal(true). So it thinks it is still in a 
transaction but dose not set the _closed.

{code}
protected void transactionComplete() {
        transactionContext = null;

        // If we were using a shared connection, clear the reference now that
        // the transaction has completed
        if (isSharedConnection) {
            setDelegate(null);
            isSharedConnection = false;
        }

        // If this connection was closed during the transaction and there is
        // still a delegate present close it
        final Connection delegate = getDelegateInternal();
        if (isClosedInternal() && delegate != null) {
            try {
                setDelegate(null);

                if (!delegate.isClosed()) {
                    delegate.close();
                }
            } catch (final SQLException ignored) {
                // Not a whole lot we can do here as connection is closed
                // and this is a transaction callback so there is no
                // way to report the error.
            }
        }
    }
{code}

The thread of transationComplete() should check isClosedInternal() which is 
still false. So it will not close the connection. And the connection.close() is 
missed both in close() and transactionComplete(). I think it could be one cause 
of the connection leaking.


> Connection leak during XATransaction in high load
> -------------------------------------------------
>
>                 Key: DBCP-484
>                 URL: https://issues.apache.org/jira/browse/DBCP-484
>             Project: Commons Dbcp
>          Issue Type: Bug
>    Affects Versions: 2.2.0
>            Reporter: Emanuel Freitas
>            Priority: Major
>
> We're experiencing a connection leak in a distributed transaction when the 
> system is under heavy load. We're using commons-dbcp (latest version) + 
> eclipselink and narayana to perform transaction coordination.
> From time to time we can see a stacktrace reporting an abandoned connection. 
> We are trying to figure out what's the root cause and we think that might be 
> some issue in the commons dbcp (not sure) . More specifically, this parte of 
> the code:
> ManagedConnection#updateTransactionStatus
> {code:java}
> if (transactionContext != null) {
>     if (transactionContext.isActive()) {
>         if (transactionContext != 
> transactionRegistry.getActiveTransactionContext()) {
>             throw new SQLException("Connection can not be used while enlisted 
> in another transaction");
>         }
>         return;
>     }
>     // transaction should have been cleared up by TransactionContextListener, 
> but in
>     // rare cases another lister could have registered which uses the 
> connection before
>     // our listener is called.  In that rare case, trigger the transaction 
> complete call now
>     transactionComplete();    
> }{code}
>  
> If the transactionContext is different than null but the state is not 
> "active" (ex: STATUS_ROLLEDBACK, STATUS_ROLLING_BACK, etc) it executes the 
> transactionComplete mothod that clears the reference to a shared connection 
> and after that the connection is never closed (returned to the pool). 
>  
> If we move the transactionComplete(); to an else,(see below), the connection 
> leak does not happen.
> {code:java}
> if (transactionContext != null) {
>     if (transactionContext.isActive()) {
>         if (transactionContext != 
> transactionRegistry.getActiveTransactionContext()) {
>             throw new SQLException("Connection can not be used while enlisted 
> in another transaction");
>         }
>         return;
>     }
> } else {
>     transactionComplete();
> }{code}
>  
> After this the dbcp unit tests still pass but I'm not sure about this 
> changes. Can you please check?
> Thanks
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to