[
https://issues.apache.org/jira/browse/TUSCANY-3475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12842627#action_12842627
]
Simon Nash commented on TUSCANY-3475:
-------------------------------------
The problem is caused by code in
org.apache.tuscany.sca.implementation.bpel.ode.EmbeddedODEServer that creates
the _executorService thread pool using the default settings of non-daemon
threads with a 60-second timeout. These non-daemon threads prevent the JVM
from exiting when the node is stopped. After 60 seconds the threads terminate
and allow the JVM to exit.
This code is based on similar code in the ODE 1.1.1 version of
org.apache.ode.axis2.ODEServer. In the latest ODE 1.3.3 version of
org.apache.ode.axis2.ODEServer this code has been changed to create a thread
pool using daemon threads and pass this thread pool into the scheduler.
I have updated the Tuscany 1.x code under r920274 so that it matches the latest
version of the ODE code. With this change the JVM exits immediately when the
node is stopped.
> Travel sample launcher for payment-bpel hangs before returning
> --------------------------------------------------------------
>
> Key: TUSCANY-3475
> URL: https://issues.apache.org/jira/browse/TUSCANY-3475
> Project: Tuscany
> Issue Type: Bug
> Components: Java SCA Travel Sample
> Reporter: Simon Nash
> Assignee: Simon Nash
>
> The travel sample payment-bpel launcher hangs for 60 seconds after the
> launcher code finishes execution and before returning back from the launcher
> main() method. This hang doesn't occur when almost identical code is run for
> the payment-bpel contribution unit test, which exits immediately when
> processing is complete. The problem occurs both when running the launcher
> from the build directory and when running the launcher from the travel sample
> binary distribution launchers directory.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
