RFR of JDK-8134599: TEST_BUG: java/rmi/transport/closeServerSocket/CloseServerSocket.java fails intermittently with Address already in use

[Roger Riggs]

Hi,

1. Replace the creation of the Registry with 
TestLibrary.crateRegistryOnEphemeralPort
and call TestLibrary.getRegistryPort to get the port number.

The intent of the test is still satisified without the timing/port 
allocation issues.

2. Another approach to testing if the port is in use would be to
assume its still active and do a registry.lookup("xxx").

The response be NotBoundException if the registry is still alive
or will be a RemoteException indicating the port is closed.

If some other application has grabbed the port, some other exception 
RemoteException cause will be thrown.

$.02, Roger


On 10/15/19 5:19 AM, Seán Coffey wrote:
> Quite a while since I touched this one! The nature of the test is a 
> bit flaky given
> that we're assuming no other process will use the chosen port. I'm not 
> sure
> if a concrete solution is possible. One other option is to have this 
> test run
> in non-concurrent mode by editing test/jdk/TEST.ROOT --> 
> exclusiveAccess.dirs=
>
> The while loop approach may help but won't protect against cases where a
> long running process has attached to the port number we're interested in.
>
> regards,
> Sean.
>
> On 15/10/2019 03:20, Weijun Wang wrote:
>> +SeanC
>>
>> The wait might unnecessarily increase the test time. Maybe you can do 
>> something like this:
>>
>>     int timeout = 10;
>>     while (timeout > 0) {
>>        sleep(one second);
>>        verifyPortFree && return;
>>        timeout--;
>>     }
>>     throw new Exception(still not freed);
>>
>> And Sean, back in JDK-8016728 you said "Let's extend it to 1000ms and 
>> see how test behaves". So what do you think?
>>
>> Thanks,
>> Max
>>
>>> On Oct 15, 2019, at 10:03 AM, Hamlin Li <huaming.li at oracle.com> wrote:
>>>
>>> Hi,
>>>
>>> The test is failing more frequently, could some help to review it?
>>>
>>> Thank you
>>>
>>> -Hamlin
>>>
>>> On 2019/9/4 11:11 AM, Hamlin Li wrote:
>>>> some background & comment: in most of failures, the 
>>>> "test.timeout.factor" is 10.0 or 8.0, so in the test code this 
>>>> factor should be considered in rmi operations such unexporting an 
>>>> object.
>>>>
>>>> Thank you
>>>>
>>>> -Hamlin
>>>>
>>>> On 2019/9/4 11:01 AM, Hamlin Li wrote:
>>>>> Hi,
>>>>>
>>>>> Would you please review the following patch?
>>>>>
>>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8134599
>>>>>
>>>>> webrev: http://cr.openjdk.java.net/~mli/8134599/webrev.00/
>>>>>
>>>>>
>>>>> Thank you
>>>>>
>>>>> -Hamlin
>>>>>