]
Richard Janík commented on WFLY-8779:
-------------------------------------
I'm cloning the issue because I'm adding a few links to upstream code and
don't want to link JBEAP issue in upstream code. (TODOs to uncomment some checks once
this is fixed).
Distributed statisics number varies among nodes (distributed
workmanager
------------------------------------------------------------------------
Key: WFLY-8779
URL:
https://issues.jboss.org/browse/WFLY-8779
Project: WildFly
Issue Type: Bug
Components: JCA
Reporter: Richard Janík
Assignee: Stefano Maestri
Reproducing test in
https://github.com/LittleJohnII/wildfly/tree/eap7-495
{noformat}
# build WFLY first
cd testsuite/integration/basic
mvn clean test -Dtest='Dwm*TestCase'
-Djboss.server.config.file.name=standalone-ha.xml -DtrimStackTrace=false
-Darquillian.launch=distributed-group
{noformat}
All the cases contain checks for distributed statistics and in some cases, these show
unexpected numbers (usually, at least one test method fails for each case). For example,
scheduling two work instances would produce distributed stats =2 for node1 and =3 for node
2.
I already mentioned this to [~maeste], so this is mainly a tracker.
This is a minor hamper to test writing for EAP7-495, but doesn't block it (not severe
enough) - if this isn't resolved before I create a PR for tests, I'll have to
comment out the distributed statistics checks.