Selection of index for CLUSTER

Hello everyone,

Is there some guideline for selection of index to use for CLUSTER? This
is rather important as physical table is reorganized according to that
index I gather? The docs are mute on this.

The docs recommend running ANALYZE on the table.

hrs=# ANALYZE VERBOSE hosts;
INFO: analyzing "public.hosts"
INFO: "hosts": scanned 47 of 47 pages, containing 765 live rows and 13
dead rows; 765 rows in sample, 765 estimated total rows
ANALYZE

Doh. Not much help here.

Regards,
mk