Re: Hot Standby (v9d)

On Sat, 2009-01-24 at 11:20 +0000, Simon Riggs wrote:
> On Sat, 2009-01-24 at 17:24 +1300, Mark Kirkwood wrote:
> 
> > > version 9g - please use this for testing now
> 
> > I'm doing some test runs with this now. I notice an old flatfiles 
> > related bug has reappeared:
> 
> I'm seeing an off-by-one error on xmax, in some cases. That then causes
> the flat file update to not pick up correct info, even though it
> executed in other ways as intended. If you run two create databases and
> then test only the first, it appears to have worked as intended.
> 
> These bugs are result of recent refactoring and it will take a few days
> to shake some of them out. We've had more than 20 already so we're
> beating them back, but we're not done yet. 

I was at a loss to explain how this could have slipped through our
tests. It appears that the error was corrected following each checkpoint
as a result of ProcArrayUpdateRunningXacts(). Our tests were performed
after a short delay, which typically would be greater than the
deliberately short setting of checkpoint_timeout/archive_timeout and so
by the time we looked the error was gone and masked the problem. We're
setting checkpoint_timeout to 30 mins now to avoid the delay...

-- Simon Riggs           www.2ndQuadrant.comPostgreSQL Training, Services and Support