05.10.2013 01:31, Alvaro Herrera kirjoitti:
> AFAICS the problem here is that this test doesn't use MultiXactIds at
> all in 9.2, but it does in 9.3. I vaguely recall Noah tried to convince
> me to put in an optimization which would have avoided this issue; I will
> give that a thought. I don't think I will be able to get it done for
> 9.3.1 though.
Is the schedule for 9.3.1 visible somewhere? Please reconsider trying
to get a fix for this in as it's a pretty big regression from 9.2. My
application test case that triggered this ran in roughly 3 seconds with
9.2 but timed out after 3 minutes on 9.3.0. And even without a loop
that amplifies this, just a single 'select for update' inside a
subtransaction is significantly slower in 9.3.
Thanks,
Oskari