Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Bruce Momjian <bruce@momjian.us> writes:
>> The idea that thousands of Postgres installations are slower just
>> so we can occasionally debug xmin/xmax issues seems way off
>> balance to me.
>
> There's no evidence whatsoever that the scope of the problem is
> that large.
Well, are we agreed that the current approach means that insertion
of a heap tuple normally requires it to be written to disk three
times, with two of those WAL-logged? And that deletion of a tuple
generally requires the same? I'd say that constitutes prima facie
evidence that any PostgreSQL installation doing any significant
number of writes is slower because of this. Are you suggesting
there aren't thousands of such installations, or that the repeated
disk writes are generally free?
-Kevin