[hibernate-dev] Hibernate ORM 5.0 test failures
sanne at hibernate.org
Tue Jun 13 04:21:51 EDT 2017
Byteman needs to open a socket into the JVM to inject its magic: you
can't otherwise rewrite the bytecode of any class.
Since various tools we use, like Byteman, Arquillian, Infinispan,
WildFly, JGroups, and probably some others I'm forgetting require
network ports we normally run CI build jobs in isolated machines.
So when you see a "BindException: Address already in use" it's most
likely caused by configuration issue of the build job not following
A rare alternative explanation is that some other phase of your same
build was not shut down correctly and is still holding on the port,
this might require more careful handling of resources.
Restarting the job might help to figure out in which case you are.
On 13 June 2017 at 07:53, Vlad Mihalcea <mihalcea.vlad at gmail.com> wrote:
> This error:
> "unexpected exception opening server socket java.net.BindException: Address
> already in use (Bind failed)"
> I only got it when NVidia driver was stealing the Wildfly port, but that
> was never for CriteriaLockingTest.
> It's also curious why the Byteman agent tries to open a socket connection:
> at org.jboss.byteman.agent.TransformListener.initialize(TransformListener.java:69)
> On Tue, Jun 13, 2017 at 9:45 AM, Gail Badner <gbadner at redhat.com> wrote:
>> I just noticed that there were a couple of failures in the CI tests:
>> Both failures are from timeouts due to:
>> java.net.BindException: Address already in use (Bind failed)
>> It passes locally for me.
>> This sounds vaguely familiar to me. Did this happen on other branches? If
>> so, what was the fix?
>> hibernate-dev mailing list
>> hibernate-dev at lists.jboss.org
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
More information about the hibernate-dev