[
https://issues.apache.org/jira/browse/ZOOKEEPER-3158?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
maoling updated ZOOKEEPER-3158:
-------------------------------
Description:
look at the source code in the ClientCnxnSocketNetty.connect(InetSocketAddress):
{code:java}
public void operationComplete(ChannelFuture channelFuture) throws Exception {
// this lock guarantees that channel won't be assgined after
cleanup().
connectLock.lock();
try {
//----------------------
sendThread.primeConnection();
//-----------------------
firstConnect.countDown();
LOG.info("channel is connected: {}",
channelFuture.getChannel());
} finally {
connectLock.unlock();
}
}
});
{code}
firstConnect.countDown() will not be executed where
sendThread.primeConnection() has thrown an exception,it should be put into
finally code block.
was:
look at the source code in the ClientCnxnSocketNetty.connect(InetSocketAddress):
*public void operationComplete(ChannelFuture channelFuture) throws Exception {
// this lock guarantees that channel won't be assgined after
cleanup().
connectLock.lock();
try {
//----------------------
sendThread.primeConnection();
//-----------------------
firstConnect.countDown();
LOG.info("channel is connected: {}",
channelFuture.getChannel());
} finally {
connectLock.unlock();
}
}
});*
firstConnect.countDown() will not be executed where
sendThread.primeConnection() has thrown an exception,it should be put into
finally code block.
> firstConnect.countDown() will not be executed where
> sendThread.primeConnection() has thrown an exception
> --------------------------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-3158
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3158
> Project: ZooKeeper
> Issue Type: Bug
> Components: server
> Reporter: maoling
> Priority: Trivial
>
> look at the source code in the
> ClientCnxnSocketNetty.connect(InetSocketAddress):
> {code:java}
> public void operationComplete(ChannelFuture channelFuture) throws Exception {
> // this lock guarantees that channel won't be assgined after
> cleanup().
> connectLock.lock();
> try {
> //----------------------
> sendThread.primeConnection();
> //-----------------------
> firstConnect.countDown();
> LOG.info("channel is connected: {}",
> channelFuture.getChannel());
> } finally {
> connectLock.unlock();
> }
> }
> });
> {code}
> firstConnect.countDown() will not be executed where
> sendThread.primeConnection() has thrown an exception,it should be put into
> finally code block.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
