On Wed, Oct 26, 2016 at 9:33 PM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> Besides, parallel CREATE INDEX alone will probably
> be quite effective at speeding up CLUSTER in practice, simply because
> that's often where most wall clock time is spent during a CLUSTER
> operation.
Creating all indexes in parallel could also help considerably, for the
case when there are multiple indexes, and seems far simpler.