All modules (except arquillian based ones) should use same surefire settings ---------------------------------------------------------------------------- Key: ISPN-2597 URL: https://issues.jboss.org/browse/ISPN-2597 Project: Infinispan Issue Type: Bug Components: Loaders and Stores Reporter: Galder Zamarreño Assignee: Galder Zamarreño Labels: hadoop Fix For: 5.2.0.Beta6, 5.2.0.Final i.e. https://infinispan.ci.cloudbees.com/job/Infinispan-master-JDK6-tcp/941/co... The reason is so that they can run within the same Maven JVM and hence avoid issues in continuous integration where resource consumption is highly controlled. This meant several changes in Maven modules and restructuring HBase cache store so that it can run with sensible settings. ---- Old description: The "ERROR: Maven JVM terminated unexpectedly with exit code 137" error is normally associated with JVM running of memory. HBase has 'perftest' forkMode which means that it creates a new JVM per test. This is way too over the top for what it needs. Trying to run it with forkMode=none and paralell testsuite logs these errors: https://gist.github.com/4224569 This cache store shouldn't be blocking things, so disabling the cache store for the moment to be able to get on with other things. Anyone with Hadoop knowledge?