Re: issue with synchronized_standby_slots

On Thu, 11 Sept 2025 at 12:00, Rahila Syed <rahilasyed90(at)gmail(dot)com> wrote:
>
> Hi,
>
>
>>
>>
>> BTW, we should also try to conclude on my yesterday's point as to why
>> it is okay to have the same behavior for default_tablespace and
>> default_table_access_method and not for this parameter? I am asking
>> because if we change the current behavior, tomorrow, we can get
>> complaints that one expects the old behaviour as that was similar to
>> other GUCs like default_tablespace and default_table_access_method.
>>
>
> Fair point. I haven't examined the validation of GUCs in parallel workers closely,
> but one argument for preventing parallel workers from failing due to an incorrect
> value of synchronized_standby_slots is that a select query works in this situation
> without parallel workers.
>
> Whereas, for incorrect values of default_tablespace and default_table_access_method,
> most commands would fail regardless of whether parallel workers are enabled.
>
> PFA a test for the original bug report on this thread. This applies on the v3 version of the patch
> that was shared.
>
Thanks for sharing the patch. I checked the test and it looks good to
me. But I am not sure if we should have a new file for the test. I
have added the test in the '040_standby_failover_slots_sync.pl' file
along with some other tests.
Also I have addressed the comments by Ashutosh in [1][2].

I have attached the updated v4 patch

[1]: https://www.postgresql.org/message-id/CAE9k0P%3Dx3J3nmSmYKmTkiFXTDKLxJkXFO4%2BVHJyNu01Od6CZfg%40mail.gmail.com
[2]: https://www.postgresql.org/message-id/CAE9k0P%3DOFMFCRy9aDGWZ3bt91tbB1WnzsAbzXN72iWBaGVuMrw%40mail.gmail.com

Thanks,
Shlok Kyal