Synchronization Master -> Slave

I've serious problems with slow link between continents, and twice a
week I have to manually reestablish the slave, by running the following
script:

------------------------------------------------------------------------------------------------------------------------------------------

psql -c "SELECT pg_start_backup('standby_sync', true)"
rsync --progress -a ${PGDATA}/*
root(at)server2(dot)mycompany(dot)com:/opt/pgsql/slave1/ --exclude postmaster.pid
--exclude postgresql.conf --exclude pg_hba.conf --exclude pg_log
--exclude pg_xlog
psql -c "SELECT pg_stop_backup()"

------------------------------------------------------------------------------------------------------------------------------------------

Point is not to analyze the connection issues between the continents,
but if there is a way better to synchronize these servers, and avoid to
transport 8Gb of data between them when syncronizing.
The rsync above can be optimized? Both servers are CentOS 5 with OpenVPN
established between them (this is one of the problems: the VPN keep
falling several times a day, and then finally, the sender and receiver
process just vanish in the air - and I had no time to investigate the
vanishment).

Thanks in advance for your help,

--

*Edson Carlos Ericksson Richter*
/SimKorp Informática Ltda/
Fone: (51) 3366-7964
Celular: (51)9318-9766/(51) 8585-0796
Embedded Image