As you indicated, this can be a RFE to show more detailed deployment status for each cluster instance.
And you are right, we don't abort/fail the overall deployment if the deployment failed on some of the cluster instances. The internal status will report back a "mixed result" for those cases. And we should make the result/status more clear to users too.
[Message sent by forum member 'hzhang_jn' (hzhang_jn)]
http://forums.java.net/jive/thread.jspa?messageID=237198