On 2021-May-25, Tom Lane wrote:
> Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> > The problem disappears completely if I add a sleep to the cancel query:
> > step "s1cancel" { SELECT pg_cancel_backend(pid), pg_sleep(0.01) FROM d3_pid; }
> > I suppose a 0.01 second sleep is not going to be sufficient to close the
> > problem in slower animals, but I hesitate to propose a much longer sleep
> > because this test has 18 permutations so even a one second sleep adds
> > quite a lot of (mostly useless) test runtime.
>
> Yeah ... maybe 0.1 second is the right tradeoff?
Pushed with a 0.1 sleep, and some commentary.
> Note that on slow (like CCA) animals, the extra query required by
> Noah's suggestion is likely to take more than 0.1 second.
Hmm, but the sleep is to compete with the cancelling of detach, not with
the noop query.
I tried running the test under CCA here and it didn't fail, but of
course that's not a guarantee of anything since it only completed one
iteration.
--
Álvaro Herrera Valdivia, Chile