This patch is needed for the reporting backend to aggregate Storage Analyzer data properly.

This patch introduces a new way of processing StorageAnalyzer data, by processing per-host instead
of per-disk. I.e. if one disk on one host is down one time data is collected, it will draw the
average size reported down, and only for this disk. The rest of the disks on the same host will show
the "full" value. Originally  it was reported per-disk. Later it was changed to average over the
time interval, which gave values that were not intuitive and too difficult to interpret.

There is an update for handling records with recently increased number of classifications.

There has also been major updates in the code.