Andres Freund <andres@anarazel.de> writes:
> On 2021-04-13 11:47:06 -0700, Paul Friedman wrote:
>> YES!!! This completely alleviates the bottleneck and I'm able to run the
>> queries full-throttle. Thank you SO much for your help+insight.
> Cool. And damn: I can't immediately think of a way to optimize this to
> not require this kind of hack in the future.
Maybe the same thing we do with user tables, ie not give up the lock
when we close a toast rel? As long as the internal lock counters
are 64-bit, we'd not have to worry about overflowing them.
regards, tom lane