Re: Relation extension scalability

On Tue, Mar 8, 2016 at 4:27 AM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>> Hmm. Can we drive this off of the heavyweight lock manager's idea of
>> how big the relation extension lock wait queue is, instead of adding
>> more stuff to PGPROC?
>
> One idea to make it work without adding additional stuff in PGPROC is that
> after acquiring relation extension lock, check if there is any available
> block in fsm, if it founds any block, then release the lock and proceed,
> else extend the relation by one block and then check lock's wait queue size
> or number of lock requests (nRequested) and extend the relation further in
> proportion to wait queue size and then release the lock and proceed. Here,
> I think we can check for wait queue size even before extending the relation
> by one block.
>
> The benefit of doing it with PGPROC is that there will be relatively less
> number LockAcquire calls as compare to heavyweight lock approach, which I
> think should not matter much because we are planing to extend the relation
> in proportion to wait queue size (probably wait queue size * 10).

I don't think switching relation extension from heavyweight locks to
lightweight locks is going to work. It would mean, for example, that
we lose the ability to service interrupts while extending a relation;
not to mention that we lose scalability if many relations are being
extended at once.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company