On 20/12/2023 17:33, Nathan Bossart wrote:
> On Wed, Dec 20, 2023 at 11:02:58AM +0200, Andrei Lepikhov wrote:
>> In that case, maybe change the test case to make it closer to real-life
>> usage - with locks and concurrent access (See attachment)?
>
> I'm not following why we should make this test case more complicated. It
> is only intended to test the DSM registry machinery, and setting/retrieving
> an atomic variable seems like a realistic use-case to me.
I could provide you at least two reasons here:
1. A More complicated example would be a tutorial on using the feature
correctly. It will reduce the number of questions in mailing lists.
2. Looking into existing extensions, I see that the most common case of
using a shared memory segment is maintaining some hash table or state
structure that needs at least one lock.
Try to rewrite the pg_prewarm according to this new feature, and you
will realize how difficult it is.
--
regards,
Andrei Lepikhov
Postgres Professional