I’ve had some problems with reading and writing UTF-8 from servlets on Tomcat 6 / Linux. request and response were utf-8, browser was utf-8, URIEncoding was set in server.xml on both connectors and hosts. Ins short, every known thing for me in code itself, and server configuration was utf-8.
When reading request, I’ve had to take byte array from String, and then convert that byte array into String again. When writing request I’ve had to write bytes, not String itself, in order to get proper response (otherwise I would get an exception that says some non ASCII character is not valid ISO 8859-1).
Changing the LANG environment variable is one way to solve the problem.
The official way is to set the character encoding in a sevlet filter: http://wiki.apache.org/tomcat/Tomcat/UTF-8
Some background information: http://www.crazysquirrel.com/computing/general/form-encoding.jspx