Hi, AgentM wrote:
I would imagine that multi-master synchronous replication would be fairly trivial to implement with 2PC and wal-shipping available, no?
Yes, that could be done. And AFAIK eigter pgpool or PgCluster (1) try to do sync, multi-master replication that way.
The problem is that such an implementation is very network intensive and scales very badly for writing transactions. If you're interested in a good sync, multi-master replication algorithm, you might want to have a look at what Slony-II [1] or Postgres-R [2] try to do. I also recommend reading "Don't be lazy, be consistent" of Bettina Kemme [3].
Regards Markus [1]: http://www.slony2.org/wiki/index.php?title=Documentation [2]: http://www.postgres-r.org/ [3]: http://www.cs.mcgill.ca/~kemme/papers/vldb00.html ---------------------------(end of broadcast)--------------------------- TIP 2: Don't 'kill -9' the postmaster