Could you please add a

Thread.currentThread().dumpStack();

into setExit() to find out who calls it?

thanks
Marcel


Peter Antman wrote:

>On 19 Sep, Marcel Ruff wrote:
>  
>
>>Peter Antman wrote:
>>
>>    
>>
>>>Hi,
>>>I would really need an advice here. When running the testsuite agains an
>>>XmlBlaster embedded in JBoss (the EmbeddedXmlBlaster) the whole JBoss
>>>server was suddenly shut down. As far as I can see it was some sort of
>>>runlevel change in XmlBlaster that triggered a System.exit call. Could
>>>that be true. In that case: this is not so good when running XmlBlaster
>>>in an embedded system, since it exits the complete JVM, and with it the
>>>whole server. Any ideas on how to make tge embedded version no call
>>>System.exit, ever! Any way to set the allowed runlevels or something
>>>like that?
>>>
>>>//Peter
>>> 
>>>
>>>      
>>>
>>Is this resolved with your TestGet.java change?
>>
>>Marcel
>>    
>>
>
>Nopp. Maybe this is the problem (from RequestBroker):
>
>
>   }
>   public void setExit(String exitValue) throws XmlBlasterException {
>      int val = 0;
>      try { val = Integer.parseInt(exitValue.trim()); } catch(NumberFormatException 
>e) { log.error(ME, "Invalid exit value=" + exitValue + ", expected an integer"); };
>      final int exitVal = val;
>      final long exitSleep = 2000L;
>      Timeout timeout = new Timeout("ExitTimer");
>      Timestamp timeoutHandle = timeout.addTimeoutListener(new I_Timeout() {
>            public void timeout(Object userData) {
>               log.info(ME, "Administrative exit(" + exitVal + ") after exit-timeout 
>of " + exitSleep + " millis.");
>               System.exit(exitVal);
>            }
>         },
>         exitSleep, null);
>      log.info(ME, "Administrative exit request, scheduled exit in " + exitSleep + " 
>millis with exit value=" + exitVal + ".");
>   }
>
>The log says:
>
>2002-09-19 10:24:45,290 549156 DEBUG [Default] (Thread-65:) [Sep 19, 2002 10:24:45 AM 
>INFO  RequestBroker-/node/http:80.72.2.80:3412] Administrative exit request, 
>scheduled exit in 2000 millis with exit value=0.
>2002-09-19 10:24:47,297 551163 DEBUG [Default] (ExitTimer:) [Sep 19, 2002 10:24:47 AM 
>INFO  RequestBroker-/node/http:80.72.2.80:3412] Administrative exit(0) after 
>exit-timeout of 2000 millis.
>
>
>//Peter
>  
>



Reply via email to