Re: Thoughts about NUM_BUFFER_PARTITIONS

On Mon, 19 Feb 2024 at 00:56, Tomas Vondra <tomas.vondra@enterprisedb.com> wrote:
> On 2/18/24 03:30, Li Japin wrote:
>>
>> I find it seems need to change MAX_SIMUL_LWLOCKS if we enlarge the NUM_BUFFER_PARTITIONS,
>> I didn’t find any comments to describe the relation between MAX_SIMUL_LWLOCKS and
>> NUM_BUFFER_PARTITIONS, am I missing someghing?
>
> IMHO the relationship is pretty simple - MAX_SIMUL_LWLOCKS needs to be
> higher than NUM_BUFFER_PARTITIONS, so that the backend can acquire all
> the partition locks if needed.
>

Thanks for the explanation!  Got it.

> There's other places that acquire a bunch of locks, and all of them need
> to be careful not to exceed MAX_SIMUL_LWLOCKS. For example gist has
> GIST_MAX_SPLIT_PAGES.
>
>
> regards