Chukwa HBase storage should scan for column from most recent time instead of starting time

HICC UI access REST API with a time range. i.e. Last 24 hours. Current ChukwaHBaseStore API is scanning for column names from start time to end time. With full scan disabled, then it is fetching a single row from the starting time to present the columns existed for that time period. Columne name retrieval does not work very well for cluster that has just been setup because there is no data in the specified start time, but there is often data for specified end time. I recommend to change the column name retrieval api to use end time as indicator.