RR, M, 02 7186723: TEST_BUG: Race condition in sun/management/jmxremote/startstop/JMXStartStopTest.sh

[Dmitry Samersoff]

Serguei,

Thank you for review!

My test "framework" contains set of tests each of them starts the server
application with different parameters.
  So the fact the server app doesn't start in one test is not necessary
imply that it wouldn't start in other tests. Typical problem - busy port.

But in jtreg mode we stop on first failing tests, so test exits if
the server app doesn't start.

-Dmitry



On 2012-09-13 03:22, serguei.spitsyn at oracle.com wrote:
> Dmitry,
> 
> It looks good.
> Just one question.
> 
> *JMXStartStopTest.sh:*
> 
>   63 _app_start(){
>   . . .
>   71   ${TESTJAVA}/bin/java -server $* -cp ${_testclasses} JMXStartStopDoSomething  >> ${_logname} 2>&1 &
>   72 
>   73   npid=`_get_pid`
>   74   if [ "${npid}" = "" ]
>   75   then
>   76      echo "ERROR: Test app not started"
>   77      if [ "${_jtreg}" = "yes" ]
>   78      then
>   79        exit -1
>   80      fi  
>   81   fi
>   82 }
> 
> Why exit at L79 is taken for jtreg only ?
> Would it make sense to do for non-jtreg mode as well?
> 
> 
> Thanks,
> Serguei
> 
> 
> On 9/11/12 9:14 AM, Dmitry Samersoff wrote:
>> Still looking for reviewers ...
>>
>>
>> On 2012-08-29 14:28, Dmitry Samersoff wrote:
>>> Hi Everybody,
>>>
>>> Second version - busy code rewritten.
>>>
>>> http://cr.openjdk.java.net/~dsamersoff/7186723/webrev.02/
>>>
>>> -Dmitry
>>>
>>> On 2012-08-25 19:18, Dmitry Samersoff wrote:
>>>> Hi Everybody,
>>>>
>>>> Test was changed to better support windows.
>>>>
>>>> http://cr.openjdk.java.net/~dsamersoff/7186723/webrev.01/
>>>>
>>>> -Dmitry
>>>>
>>>
>>
> 


-- 
Dmitry Samersoff
Java Hotspot development team, SPB04
* There will come soft rains ...