On Thu, Sep 30, 2021 at 8:18 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
>
> On Wed, Sep 29, 2021 at 8:55 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> >
> > On Wed, Sep 29, 2021 at 4:49 PM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> > >
> > > On Tue, Sep 28, 2021 at 7:04 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> > > >
> > > > On Tue, Sep 28, 2021 at 11:35 AM Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> > > > >
> > > > > On Tue, Sep 28, 2021 at 1:54 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> > > > > >
> > > > > > >
> > > > > > > Then, if, we proceed in this direction,
> > > > > > > the place to implement those stats
> > > > > > > would be on the LogicalRepWorker struct, instead ?
> > > > > > >
> > > > > >
> > > > > > Or, we can make existing stats persistent and then add these stats on
> > > > > > top of it. Sawada-San, do you have any thoughts on this matter?
> > > > >
> > > > > I think that making existing stats including received_lsn and
> > > > > last_msg_receipt_time persistent by using stats collector could cause
> > > > > massive reporting messages. We can report these messages with a
> > > > > certain interval to reduce the amount of messages but we will end up
> > > > > seeing old stats on the view.
> > > > >
> > > >
> > > > Can't we keep the current and new stats both in-memory and persist on
> > > > disk? So, the persistent stats data will be used to fill the in-memory
> > > > counters after restarting of workers, otherwise, we will always refer
> > > > to in-memory values.
> > >
> > > Interesting. Probably we can have apply workers and table sync workers
> > > send their statistics to the stats collector at exit (before the stats
> > > collector shutting down)? And the startup process will restore them at
> > > restart?
> > >
> >
> > I think we do need to send at the exit but we should probably send at
> > some other regular interval as well to avoid losing all the stats
> > after the crash-restart situation.
>
> But we clear all statistics collected by the stats collector during
> crash recovery.
>
Right. So, maybe if we do what you are suggesting is sufficient.
--
With Regards,
Amit Kapila.