[Pgpool-general] Question about clusters and replicators.

[Iñigo Martinez Lasala]

Hi everybody.
 
In order to increase our database availability and performace, we are
planning to deploy a cluster-based solution.
At this moment we have tested with several approaches, none of them
working for our application (postgres 8.2 based)
 
pgpool-II -> It does work, but we have stored procedures that use random
functions. So, due to it's a statement based solution it does not work for
our application. Obviously, we cannot change that stored procedures at
this moment.
pgcluster (1.7) -> It compiles. However, while performing a database
restore, it randomly crashes.
cybercluster (1.0.2) -> It compiles. But we have the same problem that
with pgcluster.
 
One approach we are evaluating is a "mixed" pgpool-II acting as a simple
load balancer and some kind of syncronous replica (Slony does not appear
to be a valid solution for this).
 
Does anybody have some tips about what would be the best approach to get a
HA/HP solution with postgres?
 
Thanks in advance.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pgfoundry.org/pipermail/pgpool-general/attachments/20090121/97a14a45/attachment.html