The batch subsystem does not marshall the configuration correct in some circumstances

I use a batch script to configure a domain ( https://github.com/wfink/quickstart/blob/ejb-multi-server-transfer/ejb-multi-server/install-domain.cli ) You can run that script with "jboss-cli.sh -c --file=install-domain.cli" against a fresh copy of wildfly Beta2 (upstream) The batch is running without messages, but the domain.xml looks like this <subsystem xmlns="urn:jboss:domain:batch:1.0"> <job-repository> <in-memory/> </job-repository> </subsystem> The used commands are: /profile=default-web/subsystem=batch:add() /profile=default-web/subsystem=batch/thread-pool=batch:add(keepalive-time= {time=100,unit=milliseconds}, max-threads=10) So the thread-pool is missing in the domain.xml. But if I try to tun the second command again (after the script) I get this: [domain@localhost:9990 /] /profile=default-web/subsystem=batch/thread-pool=batch:add(keepalive-time={time=100,unit=milliseconds} , max-threads=10) { "outcome" => "failed", "failure-description" => {"domain-failure-description" => "JBAS014803: Duplicate resource [ (\"profile\" => \"default-web\"), (\"subsystem\" => \"batch\"), (\"thread-pool\" => \"batch\") ]"} , "rolled-back" => true, "response-headers" => {"process-state" => "reload-required"} } After the domain is restarted the same command will run successfully.