kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guozhang Wang <wangg...@gmail.com>
Subject Re: [VOTE] KIP-330: Add retentionPeriod in SessionBytesStoreSupplier
Date Mon, 02 Jul 2018 22:55:18 GMT
+1 from myself as well.

I'm closing this thread with the tally below:

binding +1: 3 (Damian, Matthias, myself)
non-binding +1: 3 (Ted, Bill, John)


Thanks for everyone's votes!


Guozhang

On Fri, Jun 29, 2018 at 1:15 PM, Matthias J. Sax <matthias@confluent.io>
wrote:

> Thx. That makes sense. Just want to make sure the KIP clearly covers it.
>
>
> +1 (binding)
>
>
> I am only +1 because I think that there are no users customizing a
> session store supplier. This KIP goes into 2.1 what is a minor release
> that should not contain any breaking changes. We need to be careful with
> changes like this and should try to avoid them.
>
>
> -Matthias
>
>
> On 6/29/18 1:02 PM, Guozhang Wang wrote:
> > Thanks Matthias, thanks for the comments! Ack on all of them and have
> > updated the wiki:
> >
> > https://cwiki.apache.org/confluence/display/KAFKA/KIP-
> 330:+Add+retentionPeriod+in+SessionBytesStoreSupplier
> >
> >
> > Regarding the impact of this KIP, here's what I'm thinking:
> >
> > 1. For users of DSL, with customized session store, they need to pass in
> > that SessionBytesStoreSupplier into Materialized, and hence need to
> always
> > customize that interface and hence be impacted.
> > 2. For users of PAPI, they will be addStore with a `StoreBuilder`. Note
> > that users can either customize the SessionBytesStoreSupplier and pass
> into
> > `Stores#persistentSessionStore`, or they can go directly instantiate the
> > `StoreBuilder` interface. For the latter case, they are not impacted.
> >
> > Does that make sense to you?
> >
> > Guozhang
> >
> >
> > On Fri, Jun 29, 2018 at 10:44 AM Matthias J. Sax <matthias@confluent.io>
> > wrote:
> >
> >> The KIP says:
> >>
> >>> In WindowBytesStoreSupplier, we will add:
> >>
> >> Should it be `SessionBytesStoreSupplier` ?
> >>
> >>
> >> What do you mean by
> >>
> >>> Users customizing the SessionBytesStoreSupplier should not implement
> >> this function.
> >>
> >> From my understanding, this is a breaking change for all users
> >> implementing a custom `session window store`. I would expect that there
> >> are very few users but the KIP should state this clearly as a breaking
> >> API change.
> >>
> >>
> >> Nit: the JIRA link seems to be wrong.
> >>
> >>
> >> -Matthias
> >>
> >> On 6/28/18 10:12 AM, John Roesler wrote:
> >>> +1
> >>>
> >>> On Thu, Jun 28, 2018 at 4:39 AM Damian Guy <damian.guy@gmail.com>
> wrote:
> >>>
> >>>> +1
> >>>>
> >>>> On Thu, 28 Jun 2018 at 02:16 Ted Yu <yuzhihong@gmail.com> wrote:
> >>>>
> >>>>> +1
> >>>>>
> >>>>> On Wed, Jun 27, 2018 at 4:40 PM, Bill Bejeck <bbejeck@gmail.com>
> >> wrote:
> >>>>>
> >>>>>> +1
> >>>>>>
> >>>>>> -Bill
> >>>>>>
> >>>>>> On Wed, Jun 27, 2018 at 7:39 PM Guozhang Wang <wangguoz@gmail.com>
> >>>>> wrote:
> >>>>>>
> >>>>>>> Hello folks,
> >>>>>>>
> >>>>>>> I'd like to start a voting thread on KIP-330. I've intentionally
> >>>>> skipped
> >>>>>>> the discuss phase since it is a pretty straight-forward
public API
> >>>>> change
> >>>>>>> and should actually be added since day one. The bug fix
of
> KAFKA-7071
> >>>>>>> helped us to discover this overlook.
> >>>>>>>
> >>>>>>>
> >>>>>>> --
> >>>>>>> -- Guozhang
> >>>>>>>
> >>>>>>
> >>>>>
> >>>>
> >>>
> >>
> >>
> >
>
>


-- 
-- Guozhang

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message