About replication:
I know that some people are running Postgres on top of DRBD. This is
a Linux virtual block-device driver that allows replicating complete
disks across the network.
The idea is to have Postgres running on one (active) node with all disk
write operations transparently replicated to a second node (passive, no
Postgres running here). If the first node fails, the second node will
start up Postgres which will then detect an unclean shutdown of the
database and perform recovery.
I was going to try this out myself, but due to lack of time have not, but
perhaps this option is viable.
Regards
--
Helge Bahmann <bahmann@math.tu-freiberg.de> /| \__
Network admin, systems programmer /_|____\
_/\ | __)
$ ./configure \\ \|__/__|
checking whether build environment is sane... yes \\/___/ |
checking for AIX... no (we already did this) |