RFR: 8314476: TestJstatdPortAndServer.java failed with "java.rmi.NoSuchObjectException: no such object in table" [v2]

[Mark Sheppard]

On Thu, 24 Aug 2023 14:57:47 GMT, Kevin Walls <kevinw at openjdk.org> wrote:

>> Several tests from test/jdk/sun/tools/jstatd are intermittent.
>> 
>> Port clashes when run at the same time on the same machine have been a problem.
>> The RMI error "no such object in table" can mean a reference on the RMI server has been GC'd.
>> Either way, jstatd fails to startup and and the test fails.
>> 
>> These both have jstatd reporting "Could not bind" which should be recognised and retried, as we already do for (some) port in use problems. i.e. JstatdTest.java: the isJstadReady() method (will correct that typo also) checks for "Port already in use", which can be printed by TCPTransport.java.
>> 
>> Should update its test to check for "Could not bind".
>> Should limit the retries also.
>
> Kevin Walls has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Less specific error message, could be ports or other failure..

Observations from other "rmi" test scenarios suggest there might be some issue in the Registry, requiring some "deep dive" analysis. In any case, for the jstatd test, checking for a "name binding condition" and executing a retry seems reasonable

-------------

PR Comment: https://git.openjdk.org/jdk/pull/15414#issuecomment-1692189768