JBoss.orgCommunity Documentation

Chapter 17. Detecting Dead Connections

17.1. Cleaning up Dead Connection Resources on the Server
17.1.1. Closing core sessions or JMS connections that you have failed to close
17.2. Detecting failure from the client side.
17.3. Configuring Asynchronous Connection Execution

In this section we will discuss connection time-to-live (TTL) and explain how HornetQ deals with crashed clients and clients which have exited without cleanly closing their resources.

Before a HornetQ client application exits it is considered good practice that it should close its resources in a controlled manner, using a finally block.

Here's an example of a well behaved core client application closing its session and session factory in a finally block:

ServerLocator locator = null;
ClientSessionFactory sf = null;
ClientSession session = null;

try
{
   locator = HornetQClient.createServerLocatorWithoutHA(..);

   sf = locator.createClientSessionFactory();;

   session = sf.createSession(...);
   
   ... do some stuff with the session...
}
finally
{
   if (session != null)
   {
      session.close();
   }
   
   if (sf != null)
   {
      sf.close();
   }

   if(locator != null)
   {
      locator.close();
   }
}
        

And here's an example of a well behaved JMS client application:

Connection jmsConnection = null;

try
{
   ConnectionFactory jmsConnectionFactory = HornetQJMSClient.createConnectionFactoryWithoutHA(...);

   jmsConnection = jmsConnectionFactory.createConnection();

   ... do some stuff with the connection...
}
finally
{
   if (connection != null)
   {
      connection.close();
   }
}
        

Unfortunately users don't always write well behaved applications, and sometimes clients just crash so they don't have a chance to clean up their resources!

If this occurs then it can leave server side resources, like sessions, hanging on the server. If these were not removed they would cause a resource leak on the server and over time this result in the server running out of memory or other resources.

We have to balance the requirement for cleaning up dead client resources with the fact that sometimes the network between the client and the server can fail and then come back, allowing the client to reconnect. HornetQ supports client reconnection, so we don't want to clean up "dead" server side resources too soon or this will prevent any client from reconnecting, as it won't be able to find its old sessions on the server.

HornetQ makes all of this configurable. For each ClientSessionFactory we define a connection TTL. Basically, the TTL determines how long the server will keep a connection alive in the absence of any data arriving from the client. The client will automatically send "ping" packets periodically to prevent the server from closing it down. If the server doesn't receive any packets on a connection for the connection TTL time, then it will automatically close all the sessions on the server that relate to that connection.

If you're using JMS, the connection TTL is defined by the ConnectionTTL attribute on a HornetQConnectionFactory instance, or if you're deploying JMS connection factory instances direct into JNDI on the server side, you can specify it in the xml config, using the parameter connection-ttl.

The default value for connection ttl is 60000ms, i.e. 1 minute. A value of -1 for ConnectionTTL means the server will never time out the connection on the server side.

If you do not wish clients to be able to specify their own connection TTL, you can override all values used by a global value set on the server side. This can be done by specifying the connection-ttl-override attribute in the server side configuration. The default value for connection-ttl-override is -1 which means "do not override" (i.e. let clients use their own values).

In the previous section we discussed how the client sends pings to the server and how "dead" connection resources are cleaned up by the server. There's also another reason for pinging, and that's for the client to be able to detect that the server or network has failed.

As long as the client is receiving data from the server it will consider the connection to be still alive.

If the client does not receive any packets for client-failure-check-period milliseconds then it will consider the connection failed and will either initiate failover, or call any FailureListener instances (or ExceptionListener instances if you are using JMS) depending on how it has been configured.

If you're using JMS it's defined by the ClientFailureCheckPeriod attribute on a HornetQConnectionFactory instance, or if you're deploying JMS connection factory instances direct into JNDI on the server side, you can specify it in the hornetq-jms.xml configuration file, using the parameter client-failure-check-period.

The default value for client failure check period is 30000ms, i.e. 30 seconds. A value of -1 means the client will never fail the connection on the client side if no data is received from the server. Typically this is much lower than connection TTL to allow clients to reconnect in case of transitory failure.

By default, packets received on the server side are executed on the remoting thread.

It is possible instead to use a thread from a thread pool to handle some packets so that the remoting thread is not tied up for too long. However, please note that processing operations asynchronously on another thread adds a little more latency. Please note that most short running operations are always handled on the remoting thread for performance reasons. To enable asynchronous connection execution, set the parameter async-connection-execution-enabled in hornetq-configuration.xml to true (default value is true).