On 2 May 2012 13:41, Robert Haas <robertmhaas@gmail.com> wrote:
> So on further reflection I'm thinking it may be best just to stick
> with a hard conflict for now and see what feedback we get from beta
> testers.
Which is what I was expecting y'all to conclude once you'd looked at
the task in more detail.
And I'm happy with the concept of beta being a period where we listen
to feedback, not just bugs, and decide whether further refinements are
required.
What I think is possible is to alter the conflict as it hits the
backend. If a backend has enable_indexonly = off then it wouldn't be
affected by those conflicts anyhow. So if we simply record whether we
are using an index only plan then we'll know whether to ignore it or
abort. I think we can handle that by marking the snapshot at executor
startup time. Needs a few other pushups but not that hard.
The likelihood is that SQL that uses index only won't run for long
enough to be cancelled anyway.
--
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services