The CoreEnvironmentBean config method for setting the nodeIdentifier assumes the value is encoded using the UTF-8 character set. But UTF-8 is a variable width character set so a) there is no guarantee that it will respect the length restriction on node names, and b) if the input string was encoded with a different charset then when we read the bytes using UTF-8 the resulting length is indeterminate. The proposal is to add a config method to set the node identifier bytes explicitly.