On Mon, Feb 21, 2011 at 4:35 AM, Tatsuo Ishii <ishii@postgresql.org> wrote:
>> Well, good news all round.
Hello on this thread,
I'm taking a look at replication timeout with non-blocking which would
be "nice" but not required for this patch, in my understanding. But
before that, we're going to put this patch through some exercise via
pgbench to determine two things:
* How much performance is impacted
* Does it crash?
As it will be somewhat hard to prove the durability guarantees of
commit without special heroics, unless someone can suggest a
mechanism. Expect some results Real Soon.
--
fdr