Re: Support for distributed transactions

Brian Oki (boki) wrote:
> It is unclear to me how SQL statements can be executed at remote nodes
> from a single coordinator and then use distributed two-phase commit (via
> 'prepare transaction tid' and 'commit prepared'). I worked at Oracle in
> the distributed database group and could do things like the following
> using PL/SQL, where we insert the same row into the same table on three
> different nodes, including the local one:
> insert into foo ....
> insert into foo(at)node2(dot)acme(dot)com <blocked::mailto:foo(at)node2(dot)acme(dot)com>
> ....
> insert into foo(at)node3(dot)acme(dot)com <mailto:foo(at)node3(dot)acme(dot)com> ....
> COMMIT
> This sequence will insert a row into all three tables and do the
> distributed atomic commitment.
>
> My question is this: How does PostgreSQL 8.2.5 execute DML statements
> (insert, update, delete, select) on remote nodes as part of the same
> transaction? Where is the syntax specified?

Unfortunately PostgreSQL doesn't have remote or federated queries like
that. You can't reference tables in different nodes. PostgreSQL does
support two-phase commit, but what you're looking for is much more than
that.

There is, however, a contrib module called dblink, that you can use to
submit SQL queries to remote database, but it's not integrated so that
you could do "INSERT foo(at)remotenode". By using dblink in update rules on
a view you can emulate that to some extent. But dblink won't do
two-phase commit, so you'll find that the updates to remote tables are
committed as soon as the statement finishes.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com