Hi,
On 2019-10-05 11:30:39 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > I've now disabled that portion of the test.
>
> There's still a large fraction of the buildfarm that's unhappy :-(
I pushed a second patch disabling two more permutations, I've not seen
any related failures since. It's frustrating that we at the moment seem
to not have deterministic behaviour with two sessions waiting for a
tuple lock at the moment. Doesn't look like that's new though.
It seems to me that most of the XactLockTableWait() users ought to
follow the pattern outlined at the top of README.tuplock, but
don't.
Greetings,
Andres Freund