Re: parallel regression test failure

Bruce Momjian <pgman@candle.pha.pa.us> writes:
> I run it every night and it fails 25% of the time.

When did you start seeing the problem?

I just wasted an hour running eighty-some iterations of "make check"
on two different machines/OSes/architectures.  Zero failures.  I also
eyeballed recent changes in the relcache/catcache area, which seems to
be what's unhappy, without finding anything.

I think it's up to yunz as are seeing misbehavior to roll up your
sleeves and debug the problem.  There's nothing more I can do.
        regards, tom lane