Andres Freund <andres@anarazel.de> writes:
> I agree that we can't just delay cleanup till, potentially, much later , but I
> don't think that means that we have to wait 30s for each connection,
> one-by-one.
While that behavior is theoretically possible, I've heard no field
complaints about it actually happening. I don't think we should
move heaven and earth to prevent it, barring some evidence of it
being a real-world issue.
ISTM we do need to do something about this AcquireExternalFD failure
though, because that's just self-inflicted damage. (And I think it's
likely new since we invented WaitEventSets, though perhaps I'm wrong
about that.)
regards, tom lane