Cluster has to sort the whole table (or do an index scan) and reorganize the
data in the table. Does it do this by copying the data? Is that effectively
the same as a vacuum full as far as the resulting space usage?
In 7.4 does the new cluster support change this?
--
greg