On Mon, Sep 18, 2006 at 10:31:26AM +0200, Sim Zacks wrote:
> > I think the reason you are seeing failures in the first function is
> > that the initial DELETE is a no-op so it doesn't serialize anything,
> > and then there is conflict when the two INSERTs proceed in parallel.
>
> Here is a simple, reproducible example that delete doesn't cause it to use
> serial:
I wonder if you set the transaction mode to "serializable" whether that
would make a difference. In standard read-committed the way it works
below seems to be what's expected (each transaction sees what was
committed at the time is ran).
Have a ncie day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.