After changing the locking scheme from a clustered partitioned table to RLL
I noticed, as explained in the manual, that the clustered index becomes a
separate index (not included in the table structure).
Now if you place each partition on a separate device the data are well
evenly spread over the partitions.
What bothers me here is that the index is not spread over the devices but
remains on the first one.
So I'd like to see this index also partitioned and spread over the devices
or as a second option be able to place the index on another segment.