I am afraid I cannot be of much help, since the exception is thrown
from a class authored by Oracle
(oracle.soap.transport.http.OracleSOAPHTTPConnection) for which I do
not have source. However, given the underlying socket error that is
throwing the exception, the error may be a problem in the native code
implementing java.net.Socket. What JDK version are you using for the
client? Have/can you try a different one?
On 6 Jun 2003 at 14:28, McDaniel, Joe wrote:
> Supposedly, this problem is with Apache SOAP 2.2 embedded in Oracle's
> JDeveloper and Oracle Application Server. My simple application fails if I
> call the server too frequently. Here is the scenario:
>
> 1. Build the Oracle JDeveloper SOAP tutorial (very simple service just to
> get system date).
>
> 2. In the sample client, I added a loop to call the server many times as in
> :
>
> public static void main(String[] args)
> {
> int i = 0;
> try
> {
> EmbeddedMyJavaClass1WSStub stub = new
> EmbeddedMyJavaClass1WSStub();
> // Add your own code here.
> long time = System.currentTimeMillis();
> for(i = 0; i < 10000; i++) {
> stub.getDate();
> }
> System.out.println("Elapsed time was " +
> (System.currentTimeMillis() - time));
> }
> catch(Exception ex)
> {
> System.out.println("Failed after " + i + " iterations");
> ex.printStackTrace();
> }
>
> }
> 3. This will fail after some number of interations with the following
> exception:
>
> Failed after 2298 iterations
> [SOAPException: faultCode=SOAP-ENV:IOException; msg=Address already in use:
> connect; targetException=java.net.BindException: Address already in use:
> connect]
>
> at
> oracle.soap.transport.http.OracleSOAPHTTPConnection.send(OracleSOAPHTTPConne
> ction.java:765)
>
> at org.apache.soap.rpc.Call.invoke(Call.java:261)
>
> at
> mypackage3.EmbeddedMyJavaClass1WSStub.getDate(EmbeddedMyJavaClass1WSStub.jav
> a:70)
>
> at
> mypackage3.EmbeddedMyJavaClass1WSStub.main(EmbeddedMyJavaClass1WSStub.java:3
> 9)
>
>
> If you restart the program it will -- typically -- fail almost immediately.
> If you wait some time, it will work again. If you add a sleep for about 40
> to 100 ms, then it will work continuously.
>
> Any suggestions? I can probably live with the problem since my application
> will not have thousands of calls per minute, but something is wrong
> somewhere. I have not tried this in Axis so far -- anyone know if it has
> the same problems?
>
> I have tried running this with the stand-alone OC4J with the same results.
> Changing memory has no effect.
>
> Best,
>
> Joe McDaniel
> The Institute for Genomic Research
>
Scott Nichol
Do not reply directly to this e-mail address,
as it is filtered to only receive e-mail from
specific mailing lists.
