Looking for a better approach for pacemaker, DRDB VS pgsql 9.1 standby feature.

Hi there,
I am setting up a 2 node database cluster (Master/standby) on pgsql 9.1
and it looks to me that drdb option could be easier to maintain & keep
running after several failovers on a row, specially because after a
failover operation it is not too friendly to set back up automatically
the former primary as the new standby. Besides I would not like to
automate an rsync over the same data directory as it could be destroyed
by a 'wrong' syncing .
Any comments are welcome !

Thanks in advance,
A.A.