Licpoll data is read and processed in file by file and the back-end expected
to get a complete hour input from a single file. When server and client timezones
have 30 minute timezone shift this caused the processing to "loose" half an
hour from every input file.

This update adds a new binary that is run when data is arriving and will split
licpoll data files on the server if these have a 30 minute offset (compared to 
server).

The same binary can be run manually to split already collected data. As the
installed server user run binary with the following options

  /opt/openit/bin/licpollarchivefilesplitter -s YYYY-MM-DD -e YYYY-MM-DD

E.g.

  /opt/openit/bin/licpollarchivefilesplitter -s 2012-12-01 -e 2013-03-01