On Thu, Aug 26, 2021 at 10:34:48AM -0400, Bruce Momjian wrote:
> On Thu, Aug 26, 2021 at 03:59:49PM +0200, Daniel Gustafsson wrote:
>> Agreed, in this particular case I think there is merit to the idea of enforcing
>> it in the backend.
>
> OK, works for me
Indeed, there is some history here with autovacuum. I have not been
careful enough to check that. Still, putting a check on
IsBinaryUpgrade in bgworker_should_start_now() would mean that we
still keep track of the set of bgworkers registered in shared memory.
Wouldn't it be better to block things at the source, as of
RegisterBackgroundWorker()? And that would keep track of the control
we have on bgworkers in a single place. I also think that we'd better
document something about that either in bgworker.sgml or pg_upgrade's
page.
--
Michael