We have this code
in
org.apache.logging.log4j.core.net.TcpSocketManager.TcpSocketManagerFactory.createManager(String,
FactoryData):
@Override
public TcpSocketManager createManager(final String name, final
FactoryData data) {
InetAddress inetAddress;
OutputStream os;
try {
inetAddress = InetAddress.getByName(data.host);
} catch (final UnknownHostException ex) {
LOGGER.error("Could not find address of " + data.host, ex,
ex);
return null;
}
try {
// LOG4J2-1042
final Socket socket = new Socket();
socket.connect(new InetSocketAddress(data.host, data.port),
data.connectTimeoutMillis);
os = socket.getOutputStream();
return new TcpSocketManager(name, os, socket, inetAddress,
data.host, data.port,
data.connectTimeoutMillis, data.delayMillis,
data.immediateFail, data.layout);
} catch (final IOException ex) {
LOGGER.error("TcpSocketManager (" + name + ") " + ex, ex);
os = new ByteArrayOutputStream();
}
if (data.delayMillis == 0) {
return null;
}
return new TcpSocketManager(name, os, null, inetAddress,
data.host, data.port, data.connectTimeoutMillis,
data.delayMillis, data.immediateFail, data.layout);
}
Notice:
LOGGER.error("TcpSocketManager (" + name + ") " + ex, ex);
os = new ByteArrayOutputStream();
What is the point of that? It creates an unbound stream in RAM, creating
one problem while dealing with another.
I think it would be better to not create the manager at all.
Thoughts?
Gary
--
E-Mail: [email protected] | [email protected]
Java Persistence with Hibernate, Second Edition
<http://www.manning.com/bauer3/>
JUnit in Action, Second Edition <http://www.manning.com/tahchiev/>
Spring Batch in Action <http://www.manning.com/templier/>
Blog: http://garygregory.wordpress.com
Home: http://garygregory.com/
Tweet! http://twitter.com/GaryGregory
