[hibernate-dev] testJGroupsBackend fail

[Sanne Grinovero]

Hi Nicolas,
I've improved that test to be less time-sensitive, could you checkout
latest master and give it another try please?

The warnings you see in the log are nothing to worry about, they are
related to the fact that the test is using an in-memory channel to
avoid trouble with your potentially wrong network configuration (it is
not actually sending data as UDP or TCP but looping it back in another
buffer as last element in the JGroups stack).

Sanne


On 5 March 2012 16:01, Nicolas Helleringer
<nicolas.helleringer at gmail.com> wrote:
> It fails here :
>
> at
> org.hibernate.search.test.jgroups.common.JGroupsCommonTest.testJGroupsBackend(JGroupsCommonTest.java:87
>
> Code seems pretty clear :
>
>  //get slave session
> Session s = getSlaveSession();
> Transaction tx = s.beginTransaction();
>  TShirt ts = new TShirt();
> ts.setLogo( "Boston" );
> ts.setSize( "XXL" );
>  ts.setLength( 23.4d );
> TShirt ts2 = new TShirt();
> ts2.setLogo( "Mapple leaves" );
>  ts2.setSize( "L" );
> ts2.setLength( 23.42d );
> s.persist( ts );
>  s.persist( ts2 );
> tx.commit();
>
> Thread.sleep( NETWORK_TIMEOUT );
>
> FullTextSession ftSess = Search.getFullTextSession( openSession() );
> ftSess.getTransaction().begin();
>  QueryParser parser = new QueryParser(
> TestConstants.getTargetLuceneVersion(), "id", TestConstants.stopAnalyzer );
> Query luceneQuery = parser.parse( "logo:Boston or logo:Mapple leaves" );
>  org.hibernate.Query query = ftSess.createFullTextQuery( luceneQuery );
> List result = query.list();
>
> assertEquals( 2, result.size() );
>
> Do not know why it is faling on my machine
>
> I got these in the output :
>
> 16:53:04,324 (main)  WARN NAKACK:292 - use_mcast_xmit should not be used
> because the transport (SHARED_LOOPBACK) does not support IP multicasting;
> setting use_mcast_xmit to false
>
> 16:53:04,973 (main)  WARN JGroupsBackendQueueProcessor:89 - HSEARCH000007:
> FLUSH is not present in your JGroups stack!  FLUSH is needed to ensure
> messages are not dropped while new nodes join the cluster.  Will proceed,
> but inconsistencies may arise!
>
> 16:53:06,726 (Incoming-2,8de7f79c-777e-48cc-8901-c28398e3fb91,NC_NHE-23941)
>  WARN LuceneBackendQueueProcessor:101 - HSEARCH000049: Was interrupted
> while waiting for index activity to finish. Index might be inconsistent or
> have a stale lock
>
> Hope it helps
>
> Niko
>
> 2012/3/5 Hardy Ferentschik <hardy at hibernate.org>
>
>> Hi Niko,
>>
>> can you have a look at the reason for the failure? Is there something in
>> the test execution log?
>> It might just be a timing issue or maybe a multicast configuration problem
>> on windows.
>>
>> --Hardy
>>
>>
>> On Mar 5, 2012, at 4:14 PM, Nicolas Helleringer wrote:
>>
>> > On current master state I got a :
>> >
>> > Results :
>> >
>> > Failed tests:
>> >
>> >
>> testJGroupsBackend(org.hibernate.search.test.jgroups.common.JGroupsCommonTest):
>> > expected:<1> but was:<0>
>> >
>> > Tests run: 457, Failures: 1, Errors: 0, Skipped: 1
>> >
>> > I am on windows. Any ideas ?
>> >
>> > Cheers,
>> >
>> > Niko
>> > _______________________________________________
>> > hibernate-dev mailing list
>> > hibernate-dev at lists.jboss.org
>> > https://lists.jboss.org/mailman/listinfo/hibernate-dev
>>
>>
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev