On 3/28/19 12:19 AM, Alan Bateman wrote:
On 28/03/2019 00:23, Mandy Chung wrote:
On 3/27/19 4:56 PM, Lance Andersen wrote:
Hi Mandy,
On Mar 27, 2019, at 7:23 PM, Mandy Chung <mandy.ch...@oracle.com
<mailto:mandy.ch...@oracle.com>> wrote:
Hi Lance,
Do you understand what takes so long for this test to run?
Well it is executing a lot of jar commands. I did not see anything
that sticks out in the failed logs that point to anything obvious.
One thing you could do is to instrument the test to gather the timing
statistics.
This test runs the jar command at least 50 times and I/O tends to be
slower on Windows. Lance can use the test infrastructure to run it
many times and gather timings from a range of Windows hosts to get an
idea on how long it takes. I can well believe that it's close to the
timeout on some systems so increasing the timeout should help. As
regards gathering stats then it is a TestNG test with ~25 tests so we
could use @Before/@AfterMethod to get data if needed. It will clutter
the output but okay although trying to see what the timeout this test
needs.
I am okay if this test is divided into multiple tests or multiple @runs
if appropriate. That's another approach to fix it.
Mandy
