Oups ... Managed server was not started within [60] s with Java 11 in integration tests

344 views
Skip to first unread message

Miguel Moquillon

unread,
Jul 7, 2020, 5:19:21 AM7/7/20
to WildFly

Hi,

Recently I have migrated our projects to Java 11 and JEE 8. Unfortunately, some of our integration tests using Arquillian and Wildfly 20.0.0 in a managed server mode fail randomly with the following message:

19:25:02  OpenJDK 64-Bit Server VM warning: You have loaded library /opt/wildfly-for-tests/wildfly-20.0.0.Final/standalone/tmp/vfs/deployment/deploymentc139eed9545cad0f/jna-5.3.1.jar-88c036950b2d0527/contents/com/sun/jna/linux-s390x/libjnidispatch.so which might have disabled stack guard. The VM will try to fix the stack guard now.
19:25:02  It's highly recommended that you fix the library with 'execstack -c <libfile>', or link it with '-z noexecstack'.
19:25:02  17:25:02,299 INFO  [org.apache.jackrabbit.core.RepositoryImpl] (MSC service thread 1-6) Starting repository...
19:25:02  17:25:02,360 INFO  [org.apache.jackrabbit.core.nodetype.NodeTypeRegistry] (MSC service thread 1-6) no custom node type definitions found
19:25:02  17:25:02,400 INFO  [org.infinispan.PERSISTENCE] (MSC service thread 1-1) ISPN000556: Starting user marshaller 'org.wildfly.clustering.infinispan.marshalling.jboss.JBossMarshaller'
19:25:02  17:25:02,423 INFO  [org.infinispan.CONTAINER] (MSC service thread 1-1) ISPN000128: Infinispan version: Infinispan 'Turia' 10.1.8.Final
19:25:02  17:25:02,714 INFO  [org.jboss.as.clustering.infinispan] (ServerService Thread Pool -- 85) WFLYCLINF0002: Started client-mappings cache from ejb container
19:25:59  [ERROR] Tests run: 1, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 67.472 s <<< FAILURE! - in org.silverpeas.core.notification.system.synchronous.SynchronousNotificationIT
19:25:59  [ERROR] org.silverpeas.core.notification.system.synchronous.SynchronousNotificationIT  Time elapsed: 67.471 s  <<< ERROR!
19:25:59  org.jboss.arquillian.container.spi.client.container.LifecycleException: Could not start container
19:25:59  Caused by: java.util.concurrent.TimeoutException: Managed server was not started within [60] s

Our version of Java 11 is:
openjdk version "11.0.7" 2020-04-14
OpenJDK Runtime Environment (build 11.0.7+10-post-Ubuntu-2ubuntu218.04)
OpenJDK 64-Bit Server VM (build 11.0.7+10-post-Ubuntu-2ubuntu218.04, mixed mode, sharing)

After some search in the Web, it appears the timeout is thrown by Arquillian itself when the server doesn't respond in time (here 60s). Until now, we had never such an error.
When I rollback to Java 8, there is never such error. With Java 11, whatever the Wildfly distribution (for instance Wildfly 18.0.1 and now Wildfly 20.0.0) some of our integration tests fail randomly (not always the same tests).

I don't know why the EJB cache takes randomly so long time with Java 11. For information, the projects have only few EJBs (4 MDB and 1 stateless session bean); they have mainly CDI beans and JPA entities beans.

Do you have some insights about this problem?

James Perkins

unread,
Jul 7, 2020, 12:23:40 PM7/7/20
to WildFly
You can increase the startup time check in your `arquillian.xml` configuration. I'm not sure why it would take longer with Java 11 though.

Miguel Moquillon

unread,
Jul 8, 2020, 11:06:35 AM7/8/20
to WildFly
Yes, it is what I did as a workaround of the problem. Nevertheless, this error disturbs me and makes me ask questions about the performance of Wildfly in production with Java 11.
Message has been deleted

James Perkins

unread,
Nov 30, 2020, 10:47:39 AM11/30/20
to WildFly
What version of Arquillian and WildFly Arquillian are you using?

On Monday, November 30, 2020 at 7:27:04 AM UTC-8 nell.f...@gmail.com wrote:
Hi,

I have a similar Problem with Arquillian since Wildfly 19 (where Java 11 was introduced and Jakartaee): 

before every @test annotated method within a class were started after one other.

Since Wildfly 19 the next @test is starting only after 20-30 seconde. Which make the arquillian tests really slow. (more than 300 Tests x 20 Seconds, can you imagine how much a build can take after that...)

When I change the only Wildfly Server to version 18 it works super fast. Not even 1 seconds between each @test.

Cheers,
Frédéric

Miguel Moquillon

unread,
Nov 30, 2020, 11:13:37 AM11/30/20
to wil...@googlegroups.com
Currently, Arquillian 1.6.0 and Wildfly 20.0.1.

PS: all of our builds are executed within a Docker container and managed
by Jenkins.

Miguel
> *19:25:02* OpenJDK 64-Bit Server VM warning: You have
> loaded library
> /opt/wildfly-for-tests/wildfly-20.0.0.Final/standalone/tmp/vfs/deployment/deploymentc139eed9545cad0f/jna-5.3.1.jar-88c036950b2d0527/contents/com/sun/jna/linux-s390x/libjnidispatch.so
> which might have disabled stack guard. The VM will try
> to fix the stack guard now. *19:25:02* It's highly
> recommended that you fix the library with 'execstack
> -c <libfile>', or link it with '-z noexecstack'.
> *19:25:02* 17:25:02,299 INFO
> [org.apache.jackrabbit.core.RepositoryImpl] (MSC
> service thread 1-6) Starting repository... *19:25:02*
> 17:25:02,360 INFO
> [org.apache.jackrabbit.core.nodetype.NodeTypeRegistry]
> (MSC service thread 1-6) no custom node type
> definitions found *19:25:02* 17:25:02,400 INFO
> [org.infinispan.PERSISTENCE] (MSC service thread 1-1)
> ISPN000556: Starting user marshaller
> 'org.wildfly.clustering.infinispan.marshalling.jboss.JBossMarshaller'
> *19:25:02* 17:25:02,423 INFO
> [org.infinispan.CONTAINER] (MSC service thread 1-1)
> ISPN000128: Infinispan version: Infinispan 'Turia'
> 10.1.8.Final *19:25:02* 17:25:02,714 INFO
> [org.jboss.as.clustering.infinispan] (ServerService
> Thread Pool -- 85) WFLYCLINF0002: Started
> client-mappings cache from ejb container *19:25:59*
> [ERROR] Tests run: 1, Failures: 0, Errors: 1, Skipped:
> 0, Time elapsed: 67.472 s <<< FAILURE! - in
> org.silverpeas.core.notification.system.synchronous.SynchronousNotificationIT
> *19:25:59* [ERROR]
> org.silverpeas.core.notification.system.synchronous.SynchronousNotificationIT
> Time elapsed: 67.471 s <<< ERROR! *19:25:59*
> org.jboss.arquillian.container.spi.client.container.LifecycleException:
> Could not start container *19:25:59* Caused by:
> java.util.concurrent.TimeoutException: Managed server
> was not started within [60] s Our version of Java 11
> is: openjdk version "11.0.7" 2020-04-14 OpenJDK
> Runtime Environment (build
> 11.0.7+10-post-Ubuntu-2ubuntu218.04) OpenJDK 64-Bit
> Server VM (build 11.0.7+10-post-Ubuntu-2ubuntu218.04,
> mixed mode, sharing) After some search in the Web, it
> appears the timeout is thrown by Arquillian itself
> when the server doesn't respond in time (here 60s).
> Until now, we had never such an error. When I rollback
> to Java 8, there is never such error. With Java 11,
> whatever the Wildfly distribution (for instance
> Wildfly 18.0.1 and now Wildfly 20.0.0) some of our
> integration tests fail randomly (not always the same
> tests). I don't know why the EJB cache takes randomly
> so long time with Java 11. For information, the
> projects have only few EJBs (4 MDB and 1 stateless
> session bean); they have mainly CDI beans and JPA
> entities beans. Do you have some insights about this
> problem?
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "WildFly" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/wildfly/pUj5_eT5Tus/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> wildfly+u...@googlegroups.com
> <mailto:wildfly+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/wildfly/5bcd7022-17fc-47d2-87a9-43be445ee399n%40googlegroups.com
> <https://groups.google.com/d/msgid/wildfly/5bcd7022-17fc-47d2-87a9-43be445ee399n%40googlegroups.com?utm_medium=email&utm_source=footer>.

James Perkins

unread,
Nov 30, 2020, 4:56:59 PM11/30/20
to WildFly
What version of WildFly Arquillian [1] are you using? There has been no visible issues like this in the WildFly testsuite itself which uses Arquillian 1.6.0 and WildFly Arquillian 3.0.1.Final.

Miguel Moquillon

unread,
Dec 1, 2020, 3:46:26 AM12/1/20
to wil...@googlegroups.com

James Perkins

unread,
Dec 1, 2020, 11:36:51 AM12/1/20
to WildFly
This is just speculation, but you might want to try 3.0.1.Final. That is the version what upgraded to 1.6.0.Final of Arquillian.

Szymon Klepacz

unread,
Dec 1, 2020, 4:41:50 PM12/1/20
to WildFly
Thanks for the tip!

I am facing the same issue, for Java 8. Setting up VM parameter didn't help. Great to see this post, I will give it a try tomorrow.  
Why was I using Arquillian WildFly 2.2.0? Because this version is used in wildfly-jakartaee8-with-tools. Already updated in WF22alpha.  

James Perkins

unread,
Dec 1, 2020, 4:53:34 PM12/1/20
to WildFly
Ah yes. We didn't upgrade to 3.0.1.Final and 1.6.0.Final of Arquillian until WildFly 22.

Miguel Moquillon

unread,
Dec 7, 2020, 11:36:16 AM12/7/20
to wil...@googlegroups.com

James Perkins

unread,
Dec 9, 2020, 8:39:25 PM12/9/20
to WildFly
Hmmm... ...I'll have to see when I get some time if there is a way to duplicate it. We're definitely not seeing this in the WildFly testsuite an it uses Arquillian with Java 11 on certain jobs.
Reply all
Reply to author
Forward
0 new messages