Re: Improving replay of XLOG_BTREE_VACUUM records

On 8 January 2016 at 13:36, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:

> Vladimir Borodin wrote:
> >
> > > 7 янв. 2016 г., в 5:26, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
> написал(а):
> > >
> > > On Thu, Jan 7, 2016 at 12:20 AM, Alvaro Herrera
> > > <alvherre(at)2ndquadrant(dot)com <mailto:alvherre(at)2ndquadrant(dot)com>> wrote:
>
> > >> Would you please have a look at Simon's patch, in particular verify
> > >> whether it solves the performance dip in your testing environment?
> > >>
> https://www.postgresql.org/message-id/CANP8%2BjJuyExr1HnTAdZraWsWkfc-octhug7YPtzPtJcYbyi4pA%40mail.gmail.com
> > >> (Note there's an updated patch a few emails down the thread.)
> > >>
> > >> If it seems to fix the problem for you, I think we should mark yours
> > >> rejected and just apply Simon’s.
> >
> > Ok, I’ll try this patch with my use case. Basically, it’s not so easy
> > now since I’ve partitioned that big table to not have such problems
> > but there is a way to reproduce it once again. If it helps, I agree
> > that my should be rejected in favor of the Simon’s patch because my
> > patch just reduces replication lag but Simon’s seems to remove lag at
> > all.
>
> I would agree except for the observation on toast indexes. I think
> that's an important enough use case that perhaps we should have both.
>

The exclusion of toast indexes is something we can remove also, I have
recently discovered. When we access toast data we ignore MVCC, but we still
have the toast pointer and chunkid to use for rechecking our scan results.
So a later patch will add some rechecks.

So I don't think it is worth applying this patch now. I should add that I
also had a patch that did this, posted earlier IIRC. That is not the reason
to reject this, just me pointing out that I'm effectively rejecting my own
earlier patch also.

--
Simon Riggs http://www.2ndQuadrant.com/
<http://www.2ndquadrant.com/>
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services