Josh Berkus <josh@agliodbs.com> writes:
>> Even if we had a semaphore of the sort you suggest, I doubt people would
>> want DROP DATABASE to wait indefinitely. The real question here is how
>> long is it reasonable for DROP DATABASE to wait before failing ...
> 10 to 15 seconds, I'd say. Is that going to be long enough for backends to
> release, assuming the DB isn't under extreme load?
While testing this, 10 seconds seemed too long --- more than long enough
for someone to start thinking it's broken. I settled on 5 seconds which
seemed about the edge of the threshold of pain. Our experience with the
buildfarm suggests that 1 second is usually long enough (since that's
the delay we were using in the contrib regression tests, and they don't
fail often on this), so I think it'll be all right at 5.
regards, tom lane