replicate specific partition or tablespace in postgres

Question

I am considering a scenario where different partitions/tablespace are replicated on different slaves. Is it possible to achieve this? If so, by which possible schemes.

Thanks.

Answer 1

Your question is already tagged with "slony", so which problems have you encountered with it?

One of slony's drawbacks - and that is even quoted here - is the complex setup.

So another possibility is: Replicate all your data without partitioning using something like HA-JDBC, and distribute the "read only" load in the cluster.

Depending on your scenario (new system vs. handling with and existing performance bottleneck, using ORM or not), there might be different answers.