-
Bug
-
Resolution: Fixed
-
P4
-
1.4.0, 1.4.2_08, 5.0u5
-
None
-
b45
-
x86, sparc
-
solaris_8, windows_xp
I'am using The CosNaming provider.
Calling new InitialContext().bind(...) make a new ORB instance to be created.
This instance is never cleanup. The number of ReaderThread used increases.
I provide a simple test case that shows the problem.
// Step 1: Instantiate the Hello servant
HelloImpl helloRef = new HelloImpl();
while(true) {
Context initialNamingContext = new InitialContext();
initialNamingContext.rebind("HelloService", helloRef );
initialNamingContext.close();
System.out.println("Press a key to continue. Check thread stack");
System.in.read();
}
Calling new InitialContext().bind(...) make a new ORB instance to be created.
This instance is never cleanup. The number of ReaderThread used increases.
I provide a simple test case that shows the problem.
// Step 1: Instantiate the Hello servant
HelloImpl helloRef = new HelloImpl();
while(true) {
Context initialNamingContext = new InitialContext();
initialNamingContext.rebind("HelloService", helloRef );
initialNamingContext.close();
System.out.println("Press a key to continue. Check thread stack");
System.in.read();
}
- duplicates
-
JDK-6304188 RMI-IIOP threads are not disposed even after the class loader is garbage collected.
- Closed
- relates to
-
JDK-6315820 JMX Remote API workaround for CORBA memory leak is no longer needed
- Resolved
-
JDK-4865245 JSR 160 - possible thread leak with RMI/IIOP
- Closed