James Netherton created CAMEL-11947:
---------------------------------------
Summary: Possible race condition in iec60870 producer
Key: CAMEL-11947
URL: https://issues.apache.org/jira/browse/CAMEL-11947
Project: Camel
Issue Type: Bug
Components: camel-iec60870
Affects Versions: 2.20.0
Reporter: James Netherton
I've been working on integrating the iec60870 component with WIldFly. I have an
integration test
[here|https://github.com/wildfly-extras/wildfly-camel/blob/master/itests/standalone/basic/src/test/java/org/wildfly/camel/test/iec60870/Iec60870IntegrationTest.java].
When it's executed on its own, everything works fine. When it runs with the
entire test suite, I encounter a NullPointerException in the Neoscada
[Client.writeCommand()|https://github.com/eclipse/neoscada/blob/master/protocols/org.eclipse.neoscada.protocol.iec60870.client/src/org/eclipse/neoscada/protocol/iec60870/client/Client.java#L259]
method.
It seems that the Netty channel that it's trying to write to is null and that
initialisation of the channel happens async in
[Client.handleOperationComplete()|https://github.com/eclipse/neoscada/blob/master/protocols/org.eclipse.neoscada.protocol.iec60870.client/src/org/eclipse/neoscada/protocol/iec60870/client/Client.java#L138].
When I debug, I see the channel gets initialised after the call to
writeCommand().
Not sure if there's any way to detect whether the client connection has been
established before trying to use it in the
[ClientProducer|https://github.com/apache/camel/blob/master/components/camel-iec60870/src/main/java/org/apache/camel/component/iec60870/client/ClientProducer.java#L49].
Or maybe this ticket would be better placed with the neoscada project?
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
