On Fri, Sep 30, 2022 at 8:20 PM Andres Freund <andres@anarazel.de> wrote:
> I think it'd be interesting to look at per-record-type stats between two
> equivalent workload, to see where practical workloads suffer the most
> (possibly with fpw=off, to make things more repeatable).
I would expect, and Dilip's results seem to confirm, the effect to be
pretty uniform: basically, nearly every record gets bigger by 4 bytes.
That's because most records contain at least one block reference, and
if they contain multiple block references, likely all but one will be
marked BKPBLOCK_SAME_REL, so we pay the cost just once.
Because of alignment padding, the practical effect is probably that
about half of the records get bigger by 8 bytes and the other half
don't get bigger at all. But I see no reason to believe that things
are any better or worse than that. Most interesting record types are
going to contain some kind of variable-length payload, so the chances
that a 4 byte size increase pushes you across a MAXALIGN boundary seem
to be no better or worse than fifty-fifty.
> I think it'd be an OK tradeoff to optimize WAL usage for a few of the worst to
> pay off for 56bit relfilenodes. The class of problems foreclosed is large
> enough to "waste" "improvement potential" on this.
I thought about trying to buy back some space elsewhere, and I think
that would be a reasonable approach to getting this committed if we
could find a way to do it. However, I don't see a terribly obvious way
of making it happen. Trying to do it by optimizing specific WAL record
types seems like a real pain in the neck, because there's tons of
different WAL records that all have the same problem. Trying to do it
in a generic way makes more sense, and the fact that we have 2 padding
bytes available in XLogRecord seems like a place to start looking, but
the way forward from there is not clear to me.
--
Robert Haas
EDB: http://www.enterprisedb.com
