should I bother with character encoding when Transport Client talks to
Node(s)? Say that -Dfile.encoding is setup differently on machines that run
TC and Node. Could this make any harm?
I tried to look at elasticsearch starting scripts and did not find any JVM
startup parameter relevant to encoding, thus the default system encoding is
probably used by the JVM. So I assume that ES handles this internally at
the communication binary protocol level. But I would like to double-check.