Thanks. 

Yes, failovers are definitely not desireable as switchovers. I intend to
use it though with Nagios as the monitor, triggering the event handler
to failover on hard non-OK state. Can live with a few mins of lost
records(the application should be doing inserts only) but not more than
10 mins of downtime. Any advice or comment?

Regards

On Thu, 16 Mar 2006 06:29:15 -0500, "Andrew Sullivan"
<ajs at crankycanuck.ca> said:
> On Wed, Mar 15, 2006 at 09:06:06PM -0800, Andrew Ng wrote:
> > replicated to the slave, but the sequence's last value is still not
> > updated at the slave. As such, subsequent inserts would fail with
> 
> I don't think so.  The snapshots are applied in a transaction.  So
> the setval() actually happens in the same transaction: you wouldn't
> see the data at all.  The danger, of course, is that you lose data on
> failover.  That's a known risk with failover, which is why there's
> such an emphasis on doing controlled switchover, if possible.  Don't
> wait until the machine's dead: if there's evidence of trouble on your
> origin, switch over sooner.
> 
> A
> 
> 
> -- 
> Andrew Sullivan  | ajs at 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
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general
-- 
  Andrew Ng
  ngandrew at fastmail.fm

-- 
http://www.fastmail.fm - A fast, anti-spam email service.