kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bill Bejeck <bbej...@gmail.com>
Subject Re: [DISCUSS] KIP-479: Add Materialized to Join
Date Fri, 13 Sep 2019 17:29:09 GMT
Hi All,

While working on the implementation of KIP-479, some issues came to light
that the KIP as written won't work.  I have updated the KIP with a solution
I believe will solve the original problem as well as address the
impediment to the initial approach.

This update is a significant change, so please review the updated KIP
https://cwiki.apache.org/confluence/display/KAFKA/KIP-479%3A+Add+StreamJoined+config+object+to+Join
and
provide feedback.  After we conclude the discussion, there will be a
re-vote.

Thanks!
Bill

On Wed, Jul 17, 2019 at 7:01 PM Guozhang Wang <wangguoz@gmail.com> wrote:

> Hi Bill, thanks for your explanations. I'm on board with your decision too.
>
>
> Guozhang
>
> On Wed, Jul 17, 2019 at 10:20 AM Bill Bejeck <bbejeck@gmail.com> wrote:
>
> > Thanks for the response, John.
> >
> > > If I can offer my thoughts, it seems better to just document on the
> > > Stream join javadoc for the Materialized parameter that it will not
> > > make the join result queriable. I'm not opposed to the queriable flag
> > > in general, but introducing it is a much larger consideration that has
> > > previously derailed this KIP discussion. In the interest of just
> > > closing the gap and keeping the API change small, it seems better to
> > > just go with documentation for now.
> >
> > I agree with your statement here.  IMHO the most important goal of this
> KIP
> > is to not breaking existing users and gain some consistency of the API.
> >
> > I'll update the KIP accordingly.
> >
> > -Bill
> >
> > On Tue, Jul 16, 2019 at 11:55 AM John Roesler <john@confluent.io> wrote:
> >
> > > Hi Bill,
> > >
> > > Thanks for driving this KIP toward a conclusion. I'm on board with
> > > your decision.
> > >
> > > You didn't mention whether you're still proposing to add the
> > > "queriable" flag to the Materialized config object, or just document
> > > that a Stream join is never queriable. Both options have come up
> > > earlier in the discussion, so it would be good to pin this down.
> > >
> > > If I can offer my thoughts, it seems better to just document on the
> > > Stream join javadoc for the Materialized parameter that it will not
> > > make the join result queriable. I'm not opposed to the queriable flag
> > > in general, but introducing it is a much larger consideration that has
> > > previously derailed this KIP discussion. In the interest of just
> > > closing the gap and keeping the API change small, it seems better to
> > > just go with documentation for now.
> > >
> > > Thanks again,
> > > -John
> > >
> > > On Thu, Jul 11, 2019 at 2:45 PM Bill Bejeck <bbejeck@gmail.com> wrote:
> > > >
> > > > Thanks all for the great discussion so far.
> > > >
> > > > Everyone has made excellent points, and I appreciate the detail
> > everyone
> > > > has put into their arguments.
> > > >
> > > > However, after carefully evaluating all the points made so far,
> > creating
> > > an
> > > > overload with Materialized is still my #1 option.
> > > > My reasoning for saying so is two-fold:
> > > >
> > > >    1. It's a small change, and IMHO since it's consistent with our
> > > current
> > > >    API concerning state store usage, the cognitive load on users will
> > be
> > > >    minimal.
> > > >    2. It achieves the most important goal of this KIP, namely to
> close
> > > the
> > > >    gap of naming state stores independently of the join operator
> name.
> > > >
> > > > Additionally, I agree with the points made by Matthias earlier (I
> > realize
> > > > there is some overlap here).
> > > >
> > > > >  - the main purpose of this KIP is to close the naming gap what we
> > > achieve
> > > > >  - we can allow people to use the new in-memory store
> > > > >  - we allow people to enable/disable caching
> > > > >  - we unify the API
> > > > >  - we decouple querying from naming
> > > > >  - it's a small API change
> > > >
> > > > Although it's not a perfect solution,  IMHO the positives of using
> > > > Materialize far outweigh the negatives, and from what we've discussed
> > so
> > > > far, anything we implement seems to involve an additional change down
> > the
> > > > road.
> > > >
> > > > If others are still strongly opposed to using Materialized, my other
> > > > preferences would be
> > > >
> > > >    1. Add a "withStoreName" to Joined.  Although I agree with
> Guozhang
> > > that
> > > >    having a parameter that only applies to one use-case would be
> > clumsy.
> > > >    2. Add a String overload for naming the store, but this would be
> my
> > > >    least favorite option as IMHO this seems to be a step backward
> from
> > > why we
> > > >    introduced configuration objects in the first place.
> > > >
> > > > Thanks,
> > > > Bill
> > > >
> > > > On Thu, Jun 27, 2019 at 4:45 PM Matthias J. Sax <
> matthias@confluent.io
> > >
> > > > wrote:
> > > >
> > > > > Thanks for the KIP Bill!
> > > > >
> > > > > Great discussion to far.
> > > > >
> > > > > About John's idea about querying upstream stores and don't
> > materialize
> > > a
> > > > > store: I agree with Bill that this seems to be an orthogonal
> > question,
> > > > > and it might be better to treat it as an independent optimization
> and
> > > > > exclude from this KIP.
> > > > >
> > > > > > What should be the behavior if there is no store
> > > > > > configured (e.g., if Materialized with only serdes) and querying
> is
> > > > > > enabled?
> > > > >
> > > > > IMHO, this could be an error case. If one wants to query a store,
> > they
> > > > > need to provide a name -- if you don't know the name, how would you
> > > > > actually query the store (even if it would be possible to get the
> > name
> > > > > from the `TopologyDescription`, it seems clumsy).
> > > > >
> > > > > If we don't want to throw an error, materializing seems to be the
> > right
> > > > > option, to exclude "query optimization" from this KIP. I would be
> ok
> > > > > with this option, even if it's clumsy to get the name from
> > > > > `TopologyDescription`; hence, I would prefer to treat it as an
> error.
> > > > >
> > > > > > To get back to the current behavior, users would have to
> > > > > > add a "bytes store supplier" to the Materialized to indicate
> that,
> > > > > > yes, they really want a state store there.
> > > > >
> > > > > This sound like a quite subtle semantic difference on how to use
> the
> > > > > API. Might be hard to explain to users. I would prefer to not
> > > introduce it.
> > > > >
> > > > >
> > > > >
> > > > > About Guozhang's points:
> > > > >
> > > > > 1a) That is actually a good point. However, I believe we cannot get
> > > > > around this issue easily, and it seems ok to me, to expose the
> actual
> > > > > store type we are using. (More thoughts later.)
> > > > >
> > > > > 1b) I don't see an issue with allowing users to query all stores?
> > What
> > > > > is the rational behind it? What do we gain by not allowing it?
> > > > >
> > > > > 2) While I understand what you are saying, we also want/need to
> have
> > a
> > > > > way in the PAPI to allow users adding "internal/private"
> > non-queryable
> > > > > stores to a topology. That's possible via
> > > > > `Materialized#withQueryingDisabled()`. We could also update
> > > > > `Topology#addStateStore(StoreBuilder, boolean isQueryable,
> > String...)`
> > > > > to address this. Again, I agree with Bill that the current API is
> > built
> > > > > in a certain way, and if we want to change it, it should be a
> > separate
> > > > > KIP, as it seems to be an orthogonal concern.
> > > > >
> > > > > > Instead, we just restrict KIP-307 to NOT
> > > > > > use the Joined.name for state store names and always use internal
> > > names
> > > > > as
> > > > > > well, which admittedly indeed leaves a hole of not being able to
> > > cover
> > > > > all
> > > > > > internal names here
> > > > >
> > > > > I think it's important to close this gap. Naming entities seems to
> a
> > > > > binary feature: if there is a gap, the feature is more or less
> > useless,
> > > > > rendering KIP-307 void.
> > > > >
> > > > >
> > > > >
> > > > > I like John's detailed list of required features and what
> > > > > Materialized/WindowByteStoreSuppliers offers. My take is, that
> adding
> > > > > Materialized including the required run-time checks is the best
> > option
> > > > > we have, for the following reasons:
> > > > >
> > > > >  - the main purpose of this KIP is to close the naming gap what we
> > > achieve
> > > > >  - we can allow people to use the new in-memory store
> > > > >  - we allow people to enable/disable caching
> > > > >  - we unify the API
> > > > >  - we decouple querying from naming
> > > > >  - it's a small API change
> > > > >
> > > > > Adding an overload and only passing in a name, would address the
> main
> > > > > purpose of the KIP. However, it falls short on all the other
> > "goodies".
> > > > > As you mentioned, passing in `Materialized` might not be perfect
> and
> > > > > maybe we need to deprecate is at some point; but this is also true
> > for
> > > > > passing in just a name.
> > > > >
> > > > > I am also not convinced, that a `StreamJoinStore` would resolve all
> > the
> > > > > issues. In the end, as long as we are using a `WindowedStore`
> > > > > internally, we need to expose this "implemenation detail" to users
> to
> > > > > allow them to plug in a custom store. Adding `Materialized` seem to
> > be
> > > > > the best short-term fix from my point of view.
> > > > >
> > > > >
> > > > > -Matthias
> > > > >
> > > > >
> > > > > On 6/27/19 9:56 AM, Guozhang Wang wrote:
> > > > > > Hi John,
> > > > > >
> > > > > > I actually feels better about a new interface but I'm not sure if
> > we
> > > > > would
> > > > > > need the full configuration of store / log / cache, now or in the
> > > future
> > > > > > ever for stream-stream join.
> > > > > >
> > > > > > Right now I feel that 1) we want to improve our implementation of
> > > > > > stream-stream join, and potentially also allow users to customize
> > > this
> > > > > > implementation but with a more suitable interface than the
> current
> > > > > > WindowStore interface, how to do that is less clear and
> > > execution-wise
> > > > > it's
> > > > > > (arguably..) not urgent; 2) we want to close the last gap
> > > (Stream-stream
> > > > > > join) of allowing users to specify all internal names to help on
> > > backward
> > > > > > compatibility, which is urgent.
> > > > > >
> > > > > > Therefore if we want to unblock 2) from 1) in the near term, I
> feel
> > > > > > slightly inclined to just add overload functions that takes in a
> > > store
> > > > > name
> > > > > > for stream-stream joins only -- and admittedly, in the future
> this
> > > > > function
> > > > > > maybe deprecated -- i.e. if we have to do something that we "may
> > > regret"
> > > > > in
> > > > > > the future, I'd like to pick the least intrusive option.
> > > > > >
> > > > > > About `Joined#withStoreName`: since the Joined class itself is
> also
> > > used
> > > > > in
> > > > > > other join types, I feel less comfortable to have a
> > > > > `Joined#withStoreName`
> > > > > > which is only going to be used by stream-stream join. Or maybe I
> > miss
> > > > > > something here about the "latter" case that you are referring to?
> > > > > >
> > > > > >
> > > > > >
> > > > > > Guozhang
> > > > > >
> > > > > > On Mon, Jun 24, 2019 at 12:16 PM John Roesler <john@confluent.io
> >
> > > wrote:
> > > > > >
> > > > > >> Thanks Guozhang,
> > > > > >>
> > > > > >> Yep. Maybe we can consider just exactly what the join needs:
> > > > > >>
> > > > > >>> the WindowStore<Bytes, byte[]> itself is fine, if overly broad,
> > > > > >>> since the only two methods we need are `window.put(key, value,
> > > > > >>> context().timestamp())` and `WindowStoreIterator<V2> iter =
> > > > > >>> window.fetch(key, timeFrom, timeTo)`.
> > > > > >>
> > > > > >> One "middle ground" would be to extract _this_ into a new store
> > > > > >> interface, which only supports these API calls, like
> > > > > >> StreamJoinStore<K, V>. This would give us the latitude we need
> to
> > > > > >> efficiently support the exact operation without concerning
> > ourselves
> > > > > >> with all the other things a WindowStore can do (which are
> > > unreachable
> > > > > >> for the join use case). It would also let us drop "store
> > duplicates"
> > > > > >> from the main WindowStore interface, since it only exists to
> > support
> > > > > >> the join use case.
> > > > > >>
> > > > > >> If we were to add a new StreamJoinStore interface, then it'd be
> > > > > >> straightforward how we could add also
> > > > > >> `Materialized.as(StreamJoinBytesStoreSupplier)` and use
> > > Materialized,
> > > > > >> or alternatively add the ability to set the bytes store on
> Joined.
> > > > > >>
> > > > > >> Personally, I'm kind of leaning toward the latter (and also
> doing
> > > > > >> `Joined#withStoreName`), since adding the new interface to
> > > > > >> Materialized then also pollutes the interface for its _actual_
> use
> > > > > >> case of materializing a table view. Of course, to solve the
> > > immediate
> > > > > >> problem, all we need is the store name, but we might feel better
> > > about
> > > > > >> adding the store name to Joined if we _also_ feel like in the
> > > future,
> > > > > >> we would add store/log/cache configuration to Joined as well.
> > > > > >>
> > > > > >> -John
> > > > > >>
> > > > > >> On Mon, Jun 24, 2019 at 12:56 PM Guozhang Wang <
> > wangguoz@gmail.com>
> > > > > wrote:
> > > > > >>>
> > > > > >>> Hello John,
> > > > > >>>
> > > > > >>> My main concern is exactly the first point at the bottom of
> your
> > > > > analysis
> > > > > >>> here: "* configure the bytes store". I'm not sure if using a
> > window
> > > > > bytes
> > > > > >>> store would be ideal for stream-stream windowed join; e.g. we
> > could
> > > > > >>> consider two dimensional list sorted by timestamps and then by
> > > keys to
> > > > > do
> > > > > >>> the join, whereas a windowed bytes store is basically sorted by
> > key
> > > > > >> first,
> > > > > >>> then by timestamp. If we expose the Materialized to let user
> pass
> > > in a
> > > > > >>> windowed bytes store, then we would need to change that if we
> > want
> > > to
> > > > > >>> replace it with a different implementation interface.
> > > > > >>>
> > > > > >>>
> > > > > >>> Guozhang
> > > > > >>>
> > > > > >>> On Mon, Jun 24, 2019 at 8:59 AM John Roesler <
> john@confluent.io>
> > > > > wrote:
> > > > > >>>
> > > > > >>>> Hey Guozhang and Bill,
> > > > > >>>>
> > > > > >>>> For what it's worth, I agree with you both!
> > > > > >>>>
> > > > > >>>> I think it might help the discussion to look concretely at
> what
> > > > > >>>> Materialized does:
> > > > > >>>> * set a WindowBytesStoreSupplier
> > > > > >>>> * set a name
> > > > > >>>> * set the key/value serdes
> > > > > >>>> * disable/enable/configure change-logging
> > > > > >>>> * disable/enable caching
> > > > > >>>> * configure retention
> > > > > >>>>
> > > > > >>>> Further, looking into the WindowBytesStoreSupplier, the
> > interface
> > > lets
> > > > > >> you:
> > > > > >>>> * get the segment interval
> > > > > >>>> * get the window size
> > > > > >>>> * get whether "duplicates" are enabled
> > > > > >>>> * get the retention period
> > > > > >>>> * (obviously) get a WindowStore<Bytes, byte[]>
> > > > > >>>>
> > > > > >>>> We know that Materialized isn't exactly what we need for
> stream
> > > joins,
> > > > > >>>> but we can see how close Materialized is to what we need. If
> it
> > is
> > > > > >>>> close, maybe we can use it and document the gaps, and if it is
> > not
> > > > > >>>> close, then maybe we should just add what we need to Joined.
> > > > > >>>> Stream Join's requirements for its stores:
> > > > > >>>> * a multimap store (i.e., it keeps duplicates) for storing
> > general
> > > > > >>>> (not windowed) keyed records associated with their insertion
> > > time, and
> > > > > >>>> allows efficient time-bounded lookups and also efficient
> purges
> > > of old
> > > > > >>>> data.
> > > > > >>>> ** Note, a properly configured WindowBytesStoreSupplier
> > satisfies
> > > this
> > > > > >>>> requirement, and the interface supports the queries we need to
> > > verify
> > > > > >>>> the configuration at run-time
> > > > > >>>> * set a name for the store
> > > > > >>>> * do _not_ set the serdes (they are already set in Joined)
> > > > > >>>> * logging could be configurable (set to enabled now)
> > > > > >>>> * caching could be configurable (set to enabled now)
> > > > > >>>> * do _not_ configure retention (determined by JoinWindows)
> > > > > >>>>
> > > > > >>>> So, out of six capabilities for Materialized, there are two we
> > > don't
> > > > > >>>> want (serdes and retention). These would become run-time
> checks
> > > if we
> > > > > >>>> use it.
> > > > > >>>>
> > > > > >>>> A third questionable capability is to provide a
> > > > > >>>> WindowBytesStoreSupplier. Looking at whether the
> > > > > >>>> WindowBytesStoreSupplier is the right interface for Stream
> Join:
> > > > > >>>> * configuring segment interval is fine
> > > > > >>>> * should _not_ configure window size (it's determined by
> > > JoinWindows)
> > > > > >>>> * duplicates _must_ be enabled
> > > > > >>>> * retention should be _at least_ windowSize + gracePeriod, but
> > > note
> > > > > >>>> that (unlike for Table window stores) there is no utility in
> > > having a
> > > > > >>>> longer retention time.
> > > > > >>>> * the WindowStore<Bytes, byte[]> itself is fine, if overly
> > broad,
> > > > > >>>> since the only two methods we need are `window.put(key, value,
> > > > > >>>> context().timestamp())` and `WindowStoreIterator<V2> iter =
> > > > > >>>> window.fetch(key, timeFrom, timeTo)`.
> > > > > >>>>
> > > > > >>>> Thus, flattening out the overlap for WindowBytesStoreSupplier
> > > onto the
> > > > > >>>> overlap for Materialized, we have 9 capabilities total (note
> > > retention
> > > > > >>>> is duplicated), we have 4 that we don't want:
> > > > > >>>> * do _not_ set the serdes (they are already set in Joined)
> > > > > >>>> * do _not_ configure retention (determined by JoinWindows)
> > > > > >>>> * should _not_ configure window size (it's determined by
> > > JoinWindows)
> > > > > >>>> * duplicates _must_ be enabled
> > > > > >>>>
> > > > > >>>> These gaps would have to be covered with run-time checks if we
> > > re-use
> > > > > >>>> Materialized and WindowStoreBytesStoreSupplier both. Maybe
> this
> > > sounds
> > > > > >>>> bad, but consider the other side, that we get 5 new
> capabilities
> > > we
> > > > > >>>> don't require, but are still pretty nice:
> > > > > >>>> * configure the bytes store
> > > > > >>>> * set a name for the store
> > > > > >>>> * configure caching
> > > > > >>>> * configure logging
> > > > > >>>> * configure segment interval
> > > > > >>>>
> > > > > >>>> Not sure where this nets us out, but it's food for thought.
> > > > > >>>> -John
> > > > > >>>>
> > > > > >>>> On Sun, Jun 23, 2019 at 7:52 PM Guozhang Wang <
> > wangguoz@gmail.com
> > > >
> > > > > >> wrote:
> > > > > >>>>>
> > > > > >>>>> Hi Bill,
> > > > > >>>>>
> > > > > >>>>> I think by giving a Materialized param into stream-stream
> join,
> > > it's
> > > > > >> okay
> > > > > >>>>> (though still ideal) to say "we still would not expose the
> > store
> > > for
> > > > > >>>>> queries", but it would sound a bit awkward to say "we would
> > also
> > > > > >> ignore
> > > > > >>>>> whatever the passed in store supplier but just use our
> default
> > > ones"
> > > > > >> --
> > > > > >>>>> again the concern is that, if in the future we'd want to
> change
> > > the
> > > > > >>>> default
> > > > > >>>>> implementation of join algorithm which no longer rely on a
> > window
> > > > > >> store
> > > > > >>>>> with deduping enabled, then we need to change this API again
> by
> > > > > >> changing
> > > > > >>>>> the store supplier type.
> > > > > >>>>>
> > > > > >>>>> If we do want to fill this hole for stream-stream join, I
> feel
> > > just
> > > > > >>>> adding
> > > > > >>>>> a String typed store-name would even be less future-intrusive
> > if
> > > we
> > > > > >>>> expect
> > > > > >>>>> this parameter to be modified later.
> > > > > >>>>>
> > > > > >>>>> Does that makes sense?
> > > > > >>>>>
> > > > > >>>>>
> > > > > >>>>> Guozhang
> > > > > >>>>>
> > > > > >>>>> On Sat, Jun 22, 2019 at 12:51 PM Bill Bejeck <
> > bbejeck@gmail.com>
> > > > > >> wrote:
> > > > > >>>>>
> > > > > >>>>>> Thanks for the comments John and Guozhang, I'll address each
> > > one of
> > > > > >>>> your
> > > > > >>>>>> comments in turn.
> > > > > >>>>>>
> > > > > >>>>>> John,
> > > > > >>>>>>
> > > > > >>>>>>> I'm wondering about a missing quadrant from the truth table
> > > > > >> involving
> > > > > >>>>>>> whether a Materialized is stored or not and querying is
> > > > > >>>>>>> enabled/disabled... What should be the behavior if there is
> > no
> > > > > >> store
> > > > > >>>>>>> configured (e.g., if Materialized with only serdes) and
> > > querying
> > > > > >> is
> > > > > >>>>>> enabled?
> > > > > >>>>>>
> > > > > >>>>>>> It seems we have two choices:
> > > > > >>>>>>> 1. we can force creation of a state store in this case, so
> > the
> > > > > >> store
> > > > > >>>>>>> can be used to serve the queries
> > > > > >>>>>>> 2. we can provide just a queriable view, basically letting
> IQ
> > > > > >> query
> > > > > >>>>>>> into the "KTableValueGetter", which would transparently
> > > > > >> construct the
> > > > > >>>>>>> query response by applying the operator logic to the
> upstream
> > > > > >> state
> > > > > >>>> if
> > > > > >>>>>>> the operator state isn't already stored.
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>> I agree with your assertion about a missing quadrant from
> the
> > > truth
> > > > > >>>> table.
> > > > > >>>>>> Additionally, I too like the concept of a queriable view.
> > But I
> > > > > >> think
> > > > > >>>> that
> > > > > >>>>>> goes a bit beyond the scope of this KIP and would like to
> > pursue
> > > > > >> that
> > > > > >>>>>> feature as follow-on work.  Also thinking about this KIP
> some
> > > > > >> more, I'm
> > > > > >>>>>> thinking of the changes to Materialized might be a reach as
> > > well.
> > > > > >>>>>> Separating the naming from a store and its queryable state
> > seems
> > > > > >> like a
> > > > > >>>>>> complex issue in and of itself and should be treated
> > > accordingly.
> > > > > >>>>>>
> > > > > >>>>>> So here's what I'm thinking now.  We add Materialzied to
> Join,
> > > but
> > > > > >> for
> > > > > >>>> now,
> > > > > >>>>>> we internally disable querying.  I know this breaks our
> > current
> > > > > >>>> semantic
> > > > > >>>>>> approach, but I think it's crucial that we do two things in
> > this
> > > > > >> KIP
> > > > > >>>>>>
> > > > > >>>>>>    1. Break the naming of the state stores from Joined to
> > > > > >>>> Materialized, so
> > > > > >>>>>>    the naming of state stores follows our current pattern
> and
> > > > > >> enables
> > > > > >>>>>> upgrades
> > > > > >>>>>>    from 2.3 to 2.4
> > > > > >>>>>>    2. Offer the ability to configure the state stores of the
> > > join,
> > > > > >> even
> > > > > >>>>>>    providing a different implementation (i.e. in-memory) if
> > > > > >> desired.
> > > > > >>>>>>
> > > > > >>>>>> With that in mind I'm considering changing the KIP to remove
> > the
> > > > > >>>> changes to
> > > > > >>>>>> Materialized, and we document very clearly that by
> providing a
> > > > > >>>> Materialized
> > > > > >>>>>> object with a name is only for naming the state store, hence
> > the
> > > > > >>>> changelog
> > > > > >>>>>> topics and any possible configurations of the store, but
> this
> > > store
> > > > > >>>> *will
> > > > > >>>>>> not be available for IQ.*
> > > > > >>>>>>
> > > > > >>>>>> WDYT?
> > > > > >>>>>>
> > > > > >>>>>> Guozhang,
> > > > > >>>>>>
> > > > > >>>>>>> 1. About not breaking compatibility of stream-stream join
> > > > > >>>> materialized
> > > > > >>>>>>> stores: I think this is a valid issue to tackle, but after
> > > > > >> thinking
> > > > > >>>> about
> > > > > >>>>>>> it once more I'm not sure if exposing Materialized would
> be a
> > > > > >> good
> > > > > >>>>>> solution
> > > > > >>>>>>> here. My rationles:
> > > > > >>>>>>>
> > > > > >>>>>>> 1.a For stream-stream join, our current usage of
> window-store
> > > is
> > > > > >> not
> > > > > >>>>>> ideal,
> > > > > >>>>>>> and we want to modify it in the near future to be more
> > > > > >> efficient. Not
> > > > > >>>>>>> allowing users to override such state store backend gives
> us
> > > such
> > > > > >>>> freedom
> > > > > >>>>>>> (which was also considered in the original DSL design),
> > whereas
> > > > > >>>> getting a
> > > > > >>>>>>> Materialized<WindowStore> basically kicks out that freedom
> > out
> > > > > >> of the
> > > > > >>>>>>> window.
> > > > > >>>>>>> 1.b For strema-stream join, in our original design we
> intend
> > to
> > > > > >>>> "never"
> > > > > >>>>>>> want users to query the state, since it is just for
> buffering
> > > the
> > > > > >>>>>> upcoming
> > > > > >>>>>>> records from the stream. Now I know that some users may
> > indeed
> > > > > >> want
> > > > > >>>> to
> > > > > >>>>>>> query it from the debugging perspective, but still I
> > concerned
> > > > > >> about
> > > > > >>>>>>> whether leveraging IQ for debugging purposes would be the
> > right
> > > > > >>>> solution
> > > > > >>>>>>> here. And adding Materialized object opens the door to let
> > > users
> > > > > >>>> query
> > > > > >>>>>>> about it (unless we did something intentionally to still
> > > forbids
> > > > > >> it),
> > > > > >>>>>> which
> > > > > >>>>>>> also restricts us in the future.
> > > > > >>>>>>>
> > > > > >>>>>>> 2. About the coupling between Materialized.name() and
> > > queryable:
> > > > > >>>> again I
> > > > > >>>>>>> think this is a valid issue. But I'm not sure if the
> current
> > > > > >>>>>>> "withQuerryingDisabled / Enabled" at Materialized is the
> best
> > > > > >>>> approach.
> > > > > >>>>>>> Here I think I agree with John, that generally speaking
> it's
> > > > > >> better
> > > > > >>>> be a
> > > > > >>>>>>> control function on the `KTable` itself, rather than on
> > > > > >>>> `Materialized`,
> > > > > >>>>>> so
> > > > > >>>>>>> fixing it via adding functions through `Materialized` seems
> > > not a
> > > > > >>>> natural
> > > > > >>>>>> approach either.
> > > > > >>>>>>
> > > > > >>>>>> I understand your thoughts here, and up to a point, I agree
> > with
> > > > > >> you.
> > > > > >>>>>> But concerning not providing Materialized as it may restrict
> > us
> > > in
> > > > > >> the
> > > > > >>>>>> future for delivering different implementations, I'm
> wondering
> > > if
> > > > > >> we
> > > > > >>>> are
> > > > > >>>>>> doing some premature optimization here.
> > > > > >>>>>> My rationale for saying so
> > > > > >>>>>>
> > > > > >>>>>>    1. I think the cost of not allowing the naming of state
> > > stores
> > > > > >> for
> > > > > >>>> joins
> > > > > >>>>>>    is too big of a gap to leave.   IMHO for joins to follow
> > the
> > > > > >> current
> > > > > >>>>>>    pattern of using Materialized for naming state stores
> would
> > > be
> > > > > >> what
> > > > > >>>> most
> > > > > >>>>>>    users would expect to use.  As I said in my comments
> > above, I
> > > > > >> think
> > > > > >>>> we
> > > > > >>>>>>    should *not include* the changes to Materialized and
> > enforce
> > > > > >> named
> > > > > >>>>>>    stores for joins as unavailable for IQ.
> > > > > >>>>>>    2. We'll still have the join methods available without a
> > > > > >>>> Materialized
> > > > > >>>>>>    allowing us to do something different internally if a
> > > > > >> Materialized
> > > > > >>>> is
> > > > > >>>>>> not
> > > > > >>>>>>    provided.
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>> Overall, I'm thinking maybe we should still use two stones
> > > rather
> > > > > >>>> than
> > > > > >>>>>> one
> > > > > >>>>>>> to kill these two birds, and probably for this KIP we just
> > > focus
> > > > > >> on
> > > > > >>>> 1)
> > > > > >>>>>>> above. And for that I'd like to not expose the Materialized
> > > > > >> either
> > > > > >>>> for
> > > > > >>>>>>> rationales that I've listed above. Instead, we just
> restrict
> > > > > >> KIP-307
> > > > > >>>> to
> > > > > >>>>>> NOT
> > > > > >>>>>>> use the Joined.name for state store names and always use
> > > internal
> > > > > >>>> names
> > > > > >>>>>> as
> > > > > >>>>>>> well, which admittedly indeed leaves a hole of not being
> able
> > > to
> > > > > >>>> cover
> > > > > >>>>>> all
> > > > > >>>>>>> internal names here, but now I feel this `hole` may better
> be
> > > > > >> filled
> > > > > >>>> by,
> > > > > >>>>>>> e.g. not creating changelog topics but just use the
> upstream
> > to
> > > > > >>>>>>> re-bootstrap the materialized store, more concretely: when
> > > > > >>>> materializing
> > > > > >>>>>>> the store, try to piggy-back the changelog topic on an
> > existing
> > > > > >>>> topic,
> > > > > >>>>>> e.g.
> > > > > >>>>>>> a) if the stream is coming directly from some source topic
> > > > > >> (including
> > > > > >>>>>>> repartition topic), make that as changelog topic and if it
> is
> > > > > >>>> repartition
> > > > > >>>>>>> topic change the retention / data purging policy
> necessarily
> > as
> > > > > >>>> well; b)
> > > > > >>>>>> if
> > > > > >>>>>>> the stream is coming from some stateless operators,
> delegate
> > > that
> > > > > >>>>>> stateless
> > > > > >>>>>>> operator to the parent stream similar as a); if the stream
> is
> > > > > >> coming
> > > > > >>>> from
> > > > > >>>>>> a
> > > > > >>>>>>> stream-stream join which is the only stateful operator that
> > can
> > > > > >>>> result in
> > > > > >>>>>> a
> > > > > >>>>>>> stream, consider merging the join into multi-way joins
> (yes,
> > > > > >> this is
> > > > > >>>> a
> > > > > >>>>>> very
> > > > > >>>>>>> hand-wavy thought, but the point here is that we do not try
> > to
> > > > > >>>> tackle it
> > > > > >>>>>>> now but leave it for a better solution :).
> > > > > >>>>>>
> > > > > >>>>>> I really like this idea!  I agree with you in that this
> > approach
> > > > > >> to too
> > > > > >>>>>> much for adding in this KIP, but we could pick it up later
> and
> > > > > >>>> leverage the
> > > > > >>>>>> Optimization framework to accomplish this re-use.
> > > > > >>>>>> Again, while I agree we should break the naming of join
> state
> > > > > >> stores
> > > > > >>>> from
> > > > > >>>>>> KIP-307, IMHO it's something we should fix now as it will be
> > the
> > > > > >> last
> > > > > >>>> piece
> > > > > >>>>>> we can provide to give users the ability to completely make
> > > their
> > > > > >>>>>> topologies "upgrade proof" when adding additional
> operations.
> > > > > >>>>>>
> > > > > >>>>>> Thanks again to both of you for comments and I look forward
> to
> > > > > >> hearing
> > > > > >>>> back
> > > > > >>>>>> from you.
> > > > > >>>>>>
> > > > > >>>>>> Regards,
> > > > > >>>>>> Bill
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>>
> > > > > >>>>>> On Thu, Jun 20, 2019 at 2:33 PM Guozhang Wang <
> > > wangguoz@gmail.com>
> > > > > >>>> wrote:
> > > > > >>>>>>
> > > > > >>>>>>> Hello Bill,
> > > > > >>>>>>>
> > > > > >>>>>>> Thanks for the KIP. Glad to see that we can likely shooting
> > two
> > > > > >> birds
> > > > > >>>>>> with
> > > > > >>>>>>> one stone. I have some concerns though about those "two
> > birds"
> > > > > >>>>>> themselves:
> > > > > >>>>>>>
> > > > > >>>>>>> 1. About not breaking compatibility of stream-stream join
> > > > > >>>> materialized
> > > > > >>>>>>> stores: I think this is a valid issue to tackle, but after
> > > > > >> thinking
> > > > > >>>> about
> > > > > >>>>>>> it once more I'm not sure if exposing Materialized would
> be a
> > > > > >> good
> > > > > >>>>>> solution
> > > > > >>>>>>> here. My rationles:
> > > > > >>>>>>>
> > > > > >>>>>>> 1.a For stream-stream join, our current usage of
> window-store
> > > is
> > > > > >> not
> > > > > >>>>>> ideal,
> > > > > >>>>>>> and we want to modify it in the near future to be more
> > > > > >> efficient. Not
> > > > > >>>>>>> allowing users to override such state store backend gives
> us
> > > such
> > > > > >>>> freedom
> > > > > >>>>>>> (which was also considered in the original DSL design),
> > whereas
> > > > > >>>> getting a
> > > > > >>>>>>> Materialized<WindowStore> basically kicks out that freedom
> > out
> > > > > >> of the
> > > > > >>>>>>> window.
> > > > > >>>>>>> 1.b For strema-stream join, in our original design we
> intend
> > to
> > > > > >>>> "never"
> > > > > >>>>>>> want users to query the state, since it is just for
> buffering
> > > the
> > > > > >>>>>> upcoming
> > > > > >>>>>>> records from the stream. Now I know that some users may
> > indeed
> > > > > >> want
> > > > > >>>> to
> > > > > >>>>>>> query it from the debugging perspective, but still I
> > concerned
> > > > > >> about
> > > > > >>>>>>> whether leveraging IQ for debugging purposes would be the
> > right
> > > > > >>>> solution
> > > > > >>>>>>> here. And adding Materialized object opens the door to let
> > > users
> > > > > >>>> query
> > > > > >>>>>>> about it (unless we did something intentionally to still
> > > forbids
> > > > > >> it),
> > > > > >>>>>> which
> > > > > >>>>>>> also restricts us in the future.
> > > > > >>>>>>>
> > > > > >>>>>>> 2. About the coupling between Materialized.name() and
> > > queryable:
> > > > > >>>> again I
> > > > > >>>>>>> think this is a valid issue. But I'm not sure if the
> current
> > > > > >>>>>>> "withQuerryingDisabled / Enabled" at Materialized is the
> best
> > > > > >>>> approach.
> > > > > >>>>>>> Here I think I agree with John, that generally speaking
> it's
> > > > > >> better
> > > > > >>>> be a
> > > > > >>>>>>> control function on the `KTable` itself, rather than on
> > > > > >>>> `Materialized`,
> > > > > >>>>>> so
> > > > > >>>>>>> fixing it via adding functions through `Materialized` seems
> > > not a
> > > > > >>>> natural
> > > > > >>>>>>> approach either.
> > > > > >>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>> Overall, I'm thinking maybe we should still use two stones
> > > rather
> > > > > >>>> than
> > > > > >>>>>> one
> > > > > >>>>>>> to kill these two birds, and probably for this KIP we just
> > > focus
> > > > > >> on
> > > > > >>>> 1)
> > > > > >>>>>>> above. And for that I'd like to not expose the Materialized
> > > > > >> either
> > > > > >>>> for
> > > > > >>>>>>> rationales that I've listed above. Instead, we just
> restrict
> > > > > >> KIP-307
> > > > > >>>> to
> > > > > >>>>>> NOT
> > > > > >>>>>>> use the Joined.name for state store names and always use
> > > internal
> > > > > >>>> names
> > > > > >>>>>> as
> > > > > >>>>>>> well, which admittedly indeed leaves a hole of not being
> able
> > > to
> > > > > >>>> cover
> > > > > >>>>>> all
> > > > > >>>>>>> internal names here, but now I feel this `hole` may better
> be
> > > > > >> filled
> > > > > >>>> by,
> > > > > >>>>>>> e.g. not creating changelog topics but just use the
> upstream
> > to
> > > > > >>>>>>> re-bootstrap the materialized store, more concretely: when
> > > > > >>>> materializing
> > > > > >>>>>>> the store, try to piggy-back the changelog topic on an
> > existing
> > > > > >>>> topic,
> > > > > >>>>>> e.g.
> > > > > >>>>>>> a) if the stream is coming directly from some source topic
> > > > > >> (including
> > > > > >>>>>>> repartition topic), make that as changelog topic and if it
> is
> > > > > >>>> repartition
> > > > > >>>>>>> topic change the retention / data purging policy
> necessarily
> > as
> > > > > >>>> well; b)
> > > > > >>>>>> if
> > > > > >>>>>>> the stream is coming from some stateless operators,
> delegate
> > > that
> > > > > >>>>>> stateless
> > > > > >>>>>>> operator to the parent stream similar as a); if the stream
> is
> > > > > >> coming
> > > > > >>>>>> from a
> > > > > >>>>>>> stream-stream join which is the only stateful operator that
> > can
> > > > > >>>> result
> > > > > >>>>>> in a
> > > > > >>>>>>> stream, consider merging the join into multi-way joins
> (yes,
> > > > > >> this is
> > > > > >>>> a
> > > > > >>>>>> very
> > > > > >>>>>>> hand-wavy thought, but the point here is that we do not try
> > to
> > > > > >>>> tackle it
> > > > > >>>>>>> now but leave it for a better solution :).
> > > > > >>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>> Guozhang
> > > > > >>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>> On Wed, Jun 19, 2019 at 11:41 AM John Roesler <
> > > john@confluent.io
> > > > > >>>
> > > > > >>>> wrote:
> > > > > >>>>>>>
> > > > > >>>>>>>> Hi Bill,
> > > > > >>>>>>>>
> > > > > >>>>>>>> Thanks for the KIP! Awesome job catching this unexpected
> > > > > >>>> consequence
> > > > > >>>>>>>> of the prior KIPs before it was released.
> > > > > >>>>>>>>
> > > > > >>>>>>>> The proposal looks good to me. On top of just fixing the
> > > > > >> problem,
> > > > > >>>> it
> > > > > >>>>>>>> seems to address two other pain points:
> > > > > >>>>>>>> * that naming a state store automatically causes it to
> > become
> > > > > >>>>>> queriable.
> > > > > >>>>>>>> * that there's currently no way to configure the bytes
> store
> > > > > >> for
> > > > > >>>> join
> > > > > >>>>>>>> windows.
> > > > > >>>>>>>>
> > > > > >>>>>>>> It's awesome that we can fix this issue and two others
> with
> > > one
> > > > > >>>>>> feature.
> > > > > >>>>>>>>
> > > > > >>>>>>>> I'm wondering about a missing quadrant from the truth
> table
> > > > > >>>> involving
> > > > > >>>>>>>> whether a Materialized is stored or not and querying is
> > > > > >>>>>>>> enabled/disabled... What should be the behavior if there
> is
> > no
> > > > > >>>> store
> > > > > >>>>>>>> configured (e.g., if Materialized with only serdes) and
> > > > > >> querying is
> > > > > >>>>>>>> enabled?
> > > > > >>>>>>>>
> > > > > >>>>>>>> It seems we have two choices:
> > > > > >>>>>>>> 1. we can force creation of a state store in this case, so
> > the
> > > > > >>>> store
> > > > > >>>>>>>> can be used to serve the queries
> > > > > >>>>>>>> 2. we can provide just a queriable view, basically letting
> > IQ
> > > > > >> query
> > > > > >>>>>>>> into the "KTableValueGetter", which would transparently
> > > > > >> construct
> > > > > >>>> the
> > > > > >>>>>>>> query response by applying the operator logic to the
> > upstream
> > > > > >>>> state if
> > > > > >>>>>>>> the operator state isn't already stored.
> > > > > >>>>>>>>
> > > > > >>>>>>>> Offhand, it seems like the second is actually a pretty
> > awesome
> > > > > >>>>>>>> capability. But it might have an awkward interaction with
> > the
> > > > > >>>> current
> > > > > >>>>>>>> semantics. Presently, if I provide a
> Materialized.withName,
> > it
> > > > > >>>> implies
> > > > > >>>>>>>> that querying should be enabled AND that the view should
> > > > > >> actually
> > > > > >>>> be
> > > > > >>>>>>>> stored in a state store. Under option 2 above, this
> behavior
> > > > > >> would
> > > > > >>>>>>>> change to NOT provision a state store and instead just
> > consult
> > > > > >> the
> > > > > >>>>>>>> ValueGetter. To get back to the current behavior, users
> > would
> > > > > >> have
> > > > > >>>> to
> > > > > >>>>>>>> add a "bytes store supplier" to the Materialized to
> indicate
> > > > > >> that,
> > > > > >>>>>>>> yes, they really want a state store there.
> > > > > >>>>>>>>
> > > > > >>>>>>>> Behavior changes are always kind of scary, but I think in
> > this
> > > > > >>>> case,
> > > > > >>>>>>>> it might actually be preferable. In the event where only
> the
> > > > > >> name
> > > > > >>>> is
> > > > > >>>>>>>> provided, it means that people just wanted to make the
> > > > > >> operation
> > > > > >>>>>>>> result queriable. If we automatically convert this to a
> > > > > >> non-stored
> > > > > >>>>>>>> view, then simply upgrading results in the same observable
> > > > > >> behavior
> > > > > >>>>>>>> and semantics, but a linear reduction in local storage
> > > > > >> requirements
> > > > > >>>>>>>> and disk i/o, as well as a corresponding linear reduction
> in
> > > > > >> memory
> > > > > >>>>>>>> usage both on and off heap.
> > > > > >>>>>>>>
> > > > > >>>>>>>> What do you think?
> > > > > >>>>>>>> -John
> > > > > >>>>>>>>
> > > > > >>>>>>>> On Tue, Jun 18, 2019 at 9:21 PM Bill Bejeck <
> > > bbejeck@gmail.com
> > > > > >>>
> > > > > >>>> wrote:
> > > > > >>>>>>>>>
> > > > > >>>>>>>>> All,
> > > > > >>>>>>>>>
> > > > > >>>>>>>>> I'd like to start a discussion for adding a Materialized
> > > > > >>>>>> configuration
> > > > > >>>>>>>>> object to KStream.join for naming state stores involved
> in
> > > > > >> joins.
> > > > > >>>>>>>>>
> > > > > >>>>>>>>>
> > > > > >>>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>
> > > > > >>>>
> > > > > >>
> > > > >
> > >
> >
> https://cwiki.apache.org/confluence/display/KAFKA/KIP-479%3A+Add+Materialized+to+Join
> > > > > >>>>>>>>>
> > > > > >>>>>>>>> Your comments and suggestions are welcome.
> > > > > >>>>>>>>>
> > > > > >>>>>>>>> Thanks,
> > > > > >>>>>>>>> Bill
> > > > > >>>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>>
> > > > > >>>>>>> --
> > > > > >>>>>>> -- Guozhang
> > > > > >>>>>>>
> > > > > >>>>>>
> > > > > >>>>>
> > > > > >>>>>
> > > > > >>>>> --
> > > > > >>>>> -- Guozhang
> > > > > >>>>
> > > > > >>>
> > > > > >>>
> > > > > >>> --
> > > > > >>> -- Guozhang
> > > > > >>
> > > > > >
> > > > > >
> > > > >
> > > > >
> > >
> >
>
>
> --
> -- Guozhang
>

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