On Fri, 5 Jul 2019 at 12:30, Robbie Gemmell <robbie.gemm...@gmail.com>
wrote:
> FWIW the test is passing for me on two different Fedora systems with
> OpenJDK 1.8.0_212 and Oracle JDK 1.8.0_181, but the log does suggest
> an issue which could very well be the one you point to. Especially
> since the JIRA is marked fixed in 7.1.0 and so could also explain any
> difference from the 7.1.4 release under vote if the test hasnt changed
> between the branches. Though oddly, Gordon and Rob have tested 7.0.x
> before and not reported this issue, so is it a new test?
>
So the failure is completely environment dependent. IIRC, for example, the
test passes when I'm on my UK ISP because it has a "broken" DNS resolver
that resolves unknown hostnames to their own IP address. I've never had
the issue on Fedora, only on OS X. I don't actually remember what I tested
7.0.7 on. I remembered that I often got this error, and that I thought I
fixed it, but had forgotten the exact JIRA until Alex found it :-).
>
> Robbie
>
> On Fri, 5 Jul 2019 at 11:09, Oleksandr Rudyy <oru...@gmail.com> wrote:
> >
> > Gordon,
> >
> > Thanks for the logs. Though, there is nothing interesting there which
> would
> > shed light over the failure cause.
> >
> > The log file contain a record for receiving OPEN performative, and, after
> > that, in 6 seconds the test fails.
> >
> > Rob reported QPID-8150 [1] about issues with time taken by hostname
> > resolution. There are port host aliases configured in broker
> configuration
> > which are resolving host names. On some environments it take a lot of
> time.
> > It could be potential reason for the test failure.
> >
> > What is your environment?
> >
> > You can remove the hostname alias entry from the broker test
> configuration
> > to make sure that it does not impact the test.
> >
> > The diff is below
> >
> > $ git diff
> > diff --git
> >
> a/systests/protocol-tests-amqp-1-0/src/main/resources/config-protocol-tests.json
> >
> b/systests/protocol-tests-amqp-1-0/src/main/resources/config-protocol-tests.json
> > index 764ff891a..522fd5bea 100644
> > ---
> >
> a/systests/protocol-tests-amqp-1-0/src/main/resources/config-protocol-tests.json
> > +++
> >
> b/systests/protocol-tests-amqp-1-0/src/main/resources/config-protocol-tests.json
> > @@ -47,9 +47,6 @@
> > "virtualhostaliases" : [ {
> > "name" : "defaultAlias",
> > "type" : "defaultAlias"
> > - }, {
> > - "name" : "hostnameAlias",
> > - "type" : "hostnameAlias"
> > }, {
> > "name" : "nameAlias",
> > "type" : "nameAlias"
> > @@ -64,10 +61,6 @@
> > "name" : "defaultAlias",
> > "type" : "defaultAlias",
> > "durable" : true
> > - }, {
> > - "name" : "hostnameAlias",
> > - "type" : "hostnameAlias",
> > - "durable" : true
> > }, {
> > "name" : "nameAlias",
> > "type" : "nameAlias",
> >
> >
> > I hope that the above should fix the issue.
> >
> > Kind Regards,
> > Alex
> >
> > [1] https://issues.apache.org/jira/browse/QPID-8150
> >
> >
> > On Fri, 5 Jul 2019 at 09:54, Gordon Sim <g...@redhat.com> wrote:
> >
> > > On 05/07/2019 9:26 am, Oleksandr Rudyy wrote:
> > > > Hi Gordon,
> > > >
> > > > Thanks for testing the release and reporting a test failure.
> > > >
> > > > The failed test sends open performative against non existing virtual
> host
> > > > and expects to receive an "open" back containing an error
> > > "amqp:not-found"
> > > > within 6 seconds after sending "open".
> > > > The failure occurred because "open" was not received within 6
> seconds. It
> > > > could be environmental. For example, a long GC "stop the world"
> pause can
> > > > cause it, or, environment in use could be slow.
> > >
> > > I did run all the tests a second time and again it failed on that
> > > specific test only. (Not that that necessarily rules out your
> hypothesis)
> > >
> > > > The test logs could provide more details about test execution.
> > > >
> > > > If you still have a test log file
> > > >
> > >
> ./systests/protocol-tests-amqp-1-0/target/surefire-reports/java-mms.1-0/TEST-org.apache.qpid.tests.protocol.v1_0.transport.connection.OpenTest.failOpenOnNonExistingHostname.txt,
> > > > could you please send it over?
> > >
> > > Attached
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org
> > > For additional commands, e-mail: users-h...@qpid.apache.org
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org
> For additional commands, e-mail: users-h...@qpid.apache.org
>
>
