LWN: Comments on "Core team statement on replication in PostgreSQL"

Core team statement on replication in PostgreSQL

Cyberax — Thu, 12 Jun 2008 23:47:22 +0000

Nope. This is bad news. The PostgreSQL team STILL can't figure out that we need REAL
replication for critical systems.

Asynchronous replication IS NOT suitable for high-availability systems (cue: the word
"asynchronous"). And in any case Slony is already VERY easy to install (copy a module and edit
PostgreSQL config). In comparison, all synchronous replication systems require patching
PostgreSQL source.

Core team statement on replication in PostgreSQL

szidikm — Fri, 06 Jun 2008 14:33:18 +0000

This is great news.

Core team statement on replication in PostgreSQL

dmag — Thu, 05 Jun 2008 15:47:16 +0000

Yay. Simple things should be simple. Replication is a good first step. It doesn't solve every
problem, but it's useful for scaling and fail-over.

Cleaning up the simple stuff will let people move on to the complex stuff. Currently, real
database High Availability is extremely messy. If you want your cluster to be ACID, there are
no simple answers, only trade-offs.

http://spyced.blogspot.com/2006/04/database-replication.html