Re: On "multi-master"

> On Fri, 2005-10-14 at 10:48, Andrew Sullivan wrote:
> > On Fri, Oct 14, 2005 at 11:54:19PM +0900, Tatsuo Ishii wrote:
> > > Enforcement? There would be plenty of ways to achieve that. For
> > > example, you could set pg_hba.conf so that on ly the host where pgpool
> > > is running on could connect to the host where postmaster is running
> > > on.
> >
> > That just changes the problem to someone logging in from that host.
> > (This isn't a theoretical problem, by the way; it's an objection that
> > I've heard from people.)
>
> You're users shouldn't be able to do that. If they can, you've set up
> your system wrong. Only the DBA should have access to that machine.
> This is the same kind of problem as having a user log into a slony
> replicant and issue the command "drop schema _clustername cascade" being
> a problem. It's a permission / user problem.
>
> > > Right. It's your freedom that you do not use pgpool until you think
> > > it's solid enough.
> >
> > And my managers approve it :) I should note, for the record, that
> > I'm extremely impressed with pgpool. I just think we have some room
> > to grow before we can say we have something to really compete with
> > the commercial multimaster systems.
>
> Agreed. pgpool is great. But it's not RAC (That's a good thing,
> actually :)

BTW, the reason why I myself stick with pgpool is there's no
perfect or acceptable replication solution for PostgreSQL (please do
not talk about RAC or MySQL Cluster. I hate them:-).

Slony-I seems great but after all it's an async. PGCluster is even
greater. However the performance for updation is too poor(don't get me
wrong. It's a great product. The performance for SELECT is amazing,
and has truly high-avilabilty). Maybe Slony-II is one of the hope, but
I have no idea how the performance is...
--
SRA OSS, Inc. Japan
Tatsuo Ishii