Re: [HACKERS] WIP: long transactions on hot standby feedback replica / proof of concept

On Thu, Aug 16, 2018 at 2:16 PM Alexander Korotkov
<a(dot)korotkov(at)postgrespro(dot)ru> wrote:
> On Tue, Aug 14, 2018 at 12:05 PM Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com> wrote:
> >
> > On Wed, Feb 28, 2018 at 11:24 PM, Ivan Kartyshov
> > <i(dot)kartyshov(at)postgrespro(dot)ru> wrote:
> > > The main goal of my changes is to let long read-only transactions run on
> > > replica if hot_standby_feedback is turned on.
> >
> > FWIW the idea proposed on the thread[1], which allows us to disable
> > the heap truncation by vacuum per tables, might help this issue. Since
> > the original problem on that thread is a performance problem an
> > alternative proposal would be better, but I think the way would be
> > helpful for this issue too and is simple. A downside of that idea is
> > that there is additional work for user to configure the reloption to
> > each tables.
>
> Thank you for pointing this thread. It's possible to cope this
> downside to introduce GUC which would serve as default, when reloption
> is not defined. But real downside is inability to truncate the heap.
> So, options are good if they provide workaround for users, but that
> shouldn't stop us from fixing issues around heap truncation.

So, do we have any objections to committing this?

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company