Re: On "multi-master" - Mailing list pgsql-general

From Andrew Sullivan
Subject Re: On "multi-master"
Date
Msg-id 20051014155436.GC19681@phlogiston.dyndns.org
Whole thread Raw
In response to Re: On "multi-master"  (Scott Marlowe <smarlowe@g2switchworks.com>)
Responses Re: On "multi-master"
List pgsql-general
On Fri, Oct 14, 2005 at 10:20:41AM -0500, Scott Marlowe wrote:
> USers accessing machines behind the scenes is a VERY bad idea.  It's not
> a pgpool bug, is a user bug.  :)

The problem with this glib answer is that we are talking about
systems where such a "user bug" can cost people millions of dollars.
They want the _machine_ to prevent the user bug.  That's what they
think they're buying, and my understanding is that some of the other
systems provide greater protection.

Remember, a five-nines system means five minutes of downtime, all
told, per year.  People who really need that are willing to pay for
it, because it's worth it.  Most of the time, it isn't, and most so
called five-nines systems really aren't.  (There is no way you could
really claim reliable five nines performance on the in-memory-only
MySQL system, for instance: it'd be too risky, unless you could
guarantee you'd never exceed your memory.  Who's willing to guarantee
the data set won't grow unexpectedly?)

That said, using pgpool for higher-reliability, we-checked-it-real-
good systems isn't a bad idea; on the contrary.  Just let's not
pretend it's something that it isn't really.

A

--
Andrew Sullivan  | ajs@crankycanuck.ca
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmodernism.
                --Brad Holland

pgsql-general by date:

Previous
From: "Joshua D. Drake"
Date:
Subject: PostgreSQL source inspection :)
Next
From: Scott Marlowe
Date:
Subject: Re: On "multi-master"