RFR: 8230957: [TESTBUG] containers/docker/TestJcmdWithSideCar.java sporadic failures

[Langer, Christoph]

Hi Matthias,

looks good to me, too, overall and as it is active in our test system it proves to stabilize things.

But, I agree to Bob in that you should rather define some constant like

private static int EXIT_VALUE_INITIAL = -999;

and use it to initialize and check pev.

Thanks
Christoph

> -----Original Message-----
> From: Baesken, Matthias <matthias.baesken at sap.com>
> Sent: Mittwoch, 9. Oktober 2019 09:37
> To: Bob Vandette <bob.vandette at oracle.com>
> Cc: hotspot-dev at openjdk.java.net; Langer, Christoph
> <christoph.langer at sap.com>; MIKHAILO_SELEDTSOV
> <mikhailo.seledtsov at oracle.com>
> Subject: RE: RFR: 8230957: [TESTBUG]
> containers/docker/TestJcmdWithSideCar.java sporadic failures
> 
> Hi Bob I changed the comment and  renamed the variable to   retryCount .
> -999   is just an initialization and has no special meaning .
> 
> New webrev :
> 
> http://cr.openjdk.java.net/~mbaesken/webrevs/8230957.2/
> 
> Best regards, Matthias
> 
> 
> >
> > I don’t think it’s necessary to state the obvious here:
> >
> > 59     private static final long TIME_TO_WAIT_FOR_MAIN_METHOD_START
> =
> > 50 * 1000; // milliseconds; increase 5 to 50
> >
> > Where does this -999 come from?  Is this documented anywhere?  If so
> > please define a constant.  Would this
> > work with podman?
> >
> >  237             int pev = -999;
> >
> > Could you give this a more descriptive name (retryCount maybe)?
> >
> > 238             int rt = 3;
> >
> >
> > Bob.
> >
> >
> > > On Oct 8, 2019, at 7:16 AM, Baesken, Matthias
> > <matthias.baesken at sap.com> wrote:
> > >
> > > Hello, please review the following test related change .
> > >
> > > The test containers/docker/TestJcmdWithSideCar.java has been
> > reactivated after
> > >
> > > https://bugs.openjdk.java.net/browse/JDK-8228960
> > >
> > > and we see now sporadic failures.
> > >
> > > We needed to increase/adjust the timeouts and also check the Xmx
> setting
> > going into the java calls in the docker containers (we might inherit when
> from
> > the test java opts).
> > > Also a bit more output has been added to the tests to better understand
> > potential issues .
> > >
> > >
> > > Bug/webrev :
> > >
> > > https://bugs.openjdk.java.net/browse/JDK-8230957
> > >
> > > http://cr.openjdk.java.net/~mbaesken/webrevs/8230957.1/
> > >
> > >
> > > Thanks, Matthias