Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
>>>> ["maybe-revoking-partitions", "ready-to-migrate-partitions", > >>>>>>> "unknown-but-owned-partitions"] in order to be consistent with > 3c1~3. > >>>>>>> > >>>>>> > >>>>>> Ack. Updated. > >>>>>> > >>>>>> > >>>>>>> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke > all > >>>>>>> partition upon heartbeat/commit fail. What's the gain here? Do we > >> want > >>>> to > >>>>>>> keep all partitions running at this moment, to be optimistic for > the > >>>> case > >>>>>>> when no partitions get reassigned? > >>>>>>> > >>>>>> > >>>>>> That's a good catch. When REBALANCE_IN_PROGRESS is received, we can > >> just > >>>>>> re-join the group with all the currently owned partitions encoded. > >>>> Updated. > >>>>>> > >>>>>> > >>>>>>> 3. In "Recommended Upgrade Procedure", remove extra 'those': " > The > >>>>>>> 'sticky' assignor works even those there are " > >>>>>>> > >>>>>> > >>>>>> Ack, should be `even when`. > >>>>>> > >>>>>> > >>>>>>> 4. Put two "looking into the future" into a separate category > from > >>>>>>> migration session. It seems inconsistent for readers to see this > >>>> before we > >>>>>>> finished discussion for everything. > >>>>>>> > >>>>>> > >>>>>> Ack. > >>>>>> > >>>>>> > >>>>>>> 5. Have we discussed the concern on the serialization? Could the > >> new > >>>>>>> metadata we are adding grow larger than the message size cap? > >>>>>>> > >>>>>> > >>>>>> We're completing https://issues.apache.org/jira/browse/KAFKA-7149 > >> which > >>>>>> should largely reduce the message size (will update the wiki > >>>> accordingly as > >>>>>> well). > >>>>>> > >>>>>> > >>>>>>> > >>>>>>> Boyang > >>>>>>> > >>>>>>> > >>>>>>> From: Guozhang Wang > >>>>>>> Sent: Monday, April 15, 2019 9:20 AM > >>>>>>> To: dev > >>>>>>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka > >> Streams > >>>>>>> > >>>>>>> Hello Jason, > >>>>>>> > >>>>>>> I agree with you that for range / round-robin it makes less sense > to > >> be > >>>>>>> compatible with cooperative rebalance protocol. > >>>>>>> > >>>>>>> As for StickyAssignor, however, I think it would still be possible > to > >>>> make > >>>>>>> the current implementation to be compatible with cooperative > >>>> rebalance. So > >>>>>>> after pondering on different options at hand I'm now proposing this > >>>>>>> approach as listed in the upgrade section: > >>>>>>> > >>>>>>> > >>>>>>> > >>>> > >> > https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath > >>>>>>> > >>>>>>> The idea is to let assignors specify which protocols it would work > >>>> with, > >>>>>>> associating with a different name; then the upgrade path would > >> involve > >>>> a > >>>>>>> "compatible" protocol which actually still use eager behavior while > >>>>>>> encoding two assignors if possible. In "Rejected Section" (just to > >>>> clarify > >>>>>>> I'm not finalizing it as rejected, just putting it there for now, > and > >>>> if
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
to recompile the code and >> continue. >>>>>>>> >>>>>>>> >>>>>>>> Ack, makes sense. >>>>>>>> >>>>>>>> 4. Hmm.. not sure if it will work. The main issue is that the >>>>>>>>> consumer-coordinator behavior (whether to revoke all or none at >>>>>>>>> onRebalancePrepare) is independent of the selected protocol's >>>> assignor >>>>>>>>> (eager or cooperative), so even if the assignor is selected to be >> the >>>>>>>>> old-versioned one, we will still not revoke at the >>>>>>> consumer-coordinator >>>>>>>>> layer and hence has the same risk of migrating still-owned >>>> partitions, >>>>>>>>> right? >>>>>>>> >>>>>>>> >>>>>>>> Yeah, basically we would have to push the eager/cooperative logic >> into >>>>>>> the >>>>>>>> PartitionAssignor itself and make the consumer aware of the >> rebalance >>>>>>>> protocol it is compatible with. As long as an eager protocol _could_ >>>> be >>>>>>>> selected, the consumer would have to be pessimistic and do eager >>>>>>>> revocation. But if all the assignors configured in the consumer >>>> support >>>>>>>> cooperative reassignment, then either 1) a cooperative protocol will >>>> be >>>>>>>> selected and cooperative revocation can be safely used, or 2) if the >>>>>>> rest >>>>>>>> of the group does not support it, then the consumer will simply >> fail. >>>>>>>> >>>>>>>> Another point which you raised offline and I will repeat here is >> that >>>>>>> this >>>>>>>> proposal's benefit is mostly limited to sticky assignment logic. >>>>>>> Arguably >>>>>>>> the range assignor may have some incidental stickiness, particularly >>>> if >>>>>>> the >>>>>>>> group is rebalancing for a newly created or deleted topic. For other >>>>>>> cases, >>>>>>>> the proposal is mostly additional overhead since it takes an >>>> additional >>>>>>>> rebalance and many of the partitions will move. Perhaps it doesn't >>>> make >>>>>>> as >>>>>>>> much sense to use the cooperative protocol for strategies like range >>>> and >>>>>>>> round-robin. That kind of argues in favor of pushing some of the >>>> control >>>>>>>> into the assignor itself. Maybe we would not bother creating >>>>>>>> CooperativeRange as I suggested above, but it would make sense to >>>>>>> create a >>>>>>>> cooperative version of the sticky assignment strategy. I thought we >>>>>>> might >>>>>>>> have to create a new sticky assignor anyway because I can't see how >> we >>>>>>>> would get compatible behavior mixing with the old version anyway. >>>>>>>> >>>>>>>> Thanks, >>>>>>>> Jason >>>>>>>> >>>>>>>> >>>>>>>> On Thu, Apr 11, 2019 at 5:53 PM Guozhang Wang >>>>>>> wrote: >>>>>>>> >>>>>>>>> Hello Matthias: >>>>>>>>> >>>>>>>>> Thanks for your review. >>>>>>>>> >>>>>>>>> The background section uses streams assignor as well as the >>>> consumer's >>>>>>>> own >>>>>>>>> stick assignor as examples illustrating the situation, but this KIP >>>> is >>>>>>>> for >>>>>>>>> consumer coordinator itself, and the rest of the paragraph did not >>>>>>> talk >>>>>>>>> about Streams any more. If you feel it's a bit distracted I can >>>> remove >>>>>>>>> those examples. >>>>>>>>> >>>>>>>>> 10). While working on the PR I realized that the revoked partitions >>>>
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Matthias, I've thought about that before, and the reason I did not include this as part of the KIP-429 scope is that fetcher / coordinator may get quite complicated to return non-empty data if "updateAssignmentMetadataIfNeeded" returns false in KafkaConsumer. In addition, when there's a rebalance in progress, letting consumers to process data which potentially may take longer time (in Streams for example, it is related to `max.poll.interval` config) could lead to higher chance of "partitions lost" and wasted processing work. So I've decided to still keep it as simple as is today, and admittedly from a user perspective, they may see consecutive "poll" call returning no data. I will create a JIRA ticket capturing this idea for future discussion whether we should consider this as a general optimization in consumer. Does that sound good to you? Guozhang On Wed, May 22, 2019 at 4:31 AM Matthias J. Sax wrote: > Thanks to the KIP and starting the VOTE Guozhang. I am overall +1. > > One follow up thought: the KIP does not discuss in details, how `poll()` > will behave after the change. It might actually be important to ensure > that `poll()` behavior changes to be non-blocking to allow an > application to process data from non-revoked partitions while a > rebalance is happening in the background. > > Thoughts? > > > -Matthias > > > On 5/10/19 1:10 AM, Guozhang Wang wrote: > > Hello Matthias, > > > > I'm proposing to change this behavior holistically inside > > ConsumerCoordinator actually. In other words I'm trying to piggy-back > this > > behavioral fix of KAFKA-4600 along with this KIP, and the motivation for > me > > to do this piggy-backing is that, with incremental rebalancing, there > would > > be partial affected partitions as we are not revoking every body any > more. > > > > > > Guozhang > > > > > > On Thu, May 9, 2019 at 6:21 AM Matthias J. Sax > > wrote: > > > >> Thanks Guozhang! > >> > >> The simplified upgrade path is great! > >> > >> > >> Just a clarification question about the "Rebalance Callback Error > >> Handling" -- does this change affect the `ConsumerCoordinator` only if > >> incremental rebalancing is use? Or does the behavior also change for the > >> eager rebalancing case? > >> > >> > >> -Matthias > >> > >> > >> On 5/9/19 3:37 AM, Guozhang Wang wrote: > >>> Hello all, > >>> > >>> Thanks for everyone who've shared their feedbacks for this KIP! If > >> there's > >>> no further comments I'll start the voting thread by end of tomorrow. > >>> > >>> > >>> Guozhang. > >>> > >>> On Wed, May 8, 2019 at 6:36 PM Guozhang Wang > wrote: > >>> > >>>> Hello Boyang, > >>>> > >>>> On Wed, May 1, 2019 at 4:51 PM Boyang Chen > wrote: > >>>> > >>>>> Hey Guozhang, > >>>>> > >>>>> thank you for the great write up. Overall the motivation and changes > >>>>> LGTM, just some minor comments: > >>>>> > >>>>> > >>>>> 1. In "Consumer Coordinator Algorithm", we could reorder alphabet > >>>>> points for 3d~3f from ["ready-to-migrate-partitions", > >>>>> "unknown-but-owned-partitions", "maybe-revoking-partitions"] to > >>>>> ["maybe-revoking-partitions", "ready-to-migrate-partitions", > >>>>> "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. > >>>>> > >>>> > >>>> Ack. Updated. > >>>> > >>>> > >>>>> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all > >>>>> partition upon heartbeat/commit fail. What's the gain here? Do we > want > >> to > >>>>> keep all partitions running at this moment, to be optimistic for the > >> case > >>>>> when no partitions get reassigned? > >>>>> > >>>> > >>>> That's a good catch. When REBALANCE_IN_PROGRESS is received, we can > just > >>>> re-join the group with all the currently owned partitions encoded. > >> Updated. > >>>> > >>>> > >>>>> 3. In "Recommended Upgrade Procedure", remove extra 'those': " The > >>>>> 'sticky' assignor wo
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Thanks to the KIP and starting the VOTE Guozhang. I am overall +1. One follow up thought: the KIP does not discuss in details, how `poll()` will behave after the change. It might actually be important to ensure that `poll()` behavior changes to be non-blocking to allow an application to process data from non-revoked partitions while a rebalance is happening in the background. Thoughts? -Matthias On 5/10/19 1:10 AM, Guozhang Wang wrote: > Hello Matthias, > > I'm proposing to change this behavior holistically inside > ConsumerCoordinator actually. In other words I'm trying to piggy-back this > behavioral fix of KAFKA-4600 along with this KIP, and the motivation for me > to do this piggy-backing is that, with incremental rebalancing, there would > be partial affected partitions as we are not revoking every body any more. > > > Guozhang > > > On Thu, May 9, 2019 at 6:21 AM Matthias J. Sax > wrote: > >> Thanks Guozhang! >> >> The simplified upgrade path is great! >> >> >> Just a clarification question about the "Rebalance Callback Error >> Handling" -- does this change affect the `ConsumerCoordinator` only if >> incremental rebalancing is use? Or does the behavior also change for the >> eager rebalancing case? >> >> >> -Matthias >> >> >> On 5/9/19 3:37 AM, Guozhang Wang wrote: >>> Hello all, >>> >>> Thanks for everyone who've shared their feedbacks for this KIP! If >> there's >>> no further comments I'll start the voting thread by end of tomorrow. >>> >>> >>> Guozhang. >>> >>> On Wed, May 8, 2019 at 6:36 PM Guozhang Wang wrote: >>> >>>> Hello Boyang, >>>> >>>> On Wed, May 1, 2019 at 4:51 PM Boyang Chen wrote: >>>> >>>>> Hey Guozhang, >>>>> >>>>> thank you for the great write up. Overall the motivation and changes >>>>> LGTM, just some minor comments: >>>>> >>>>> >>>>> 1. In "Consumer Coordinator Algorithm", we could reorder alphabet >>>>> points for 3d~3f from ["ready-to-migrate-partitions", >>>>> "unknown-but-owned-partitions", "maybe-revoking-partitions"] to >>>>> ["maybe-revoking-partitions", "ready-to-migrate-partitions", >>>>> "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. >>>>> >>>> >>>> Ack. Updated. >>>> >>>> >>>>> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all >>>>> partition upon heartbeat/commit fail. What's the gain here? Do we want >> to >>>>> keep all partitions running at this moment, to be optimistic for the >> case >>>>> when no partitions get reassigned? >>>>> >>>> >>>> That's a good catch. When REBALANCE_IN_PROGRESS is received, we can just >>>> re-join the group with all the currently owned partitions encoded. >> Updated. >>>> >>>> >>>>> 3. In "Recommended Upgrade Procedure", remove extra 'those': " The >>>>> 'sticky' assignor works even those there are " >>>>> >>>> >>>> Ack, should be `even when`. >>>> >>>> >>>>> 4. Put two "looking into the future" into a separate category from >>>>> migration session. It seems inconsistent for readers to see this >> before we >>>>> finished discussion for everything. >>>>> >>>> >>>> Ack. >>>> >>>> >>>>> 5. Have we discussed the concern on the serialization? Could the new >>>>> metadata we are adding grow larger than the message size cap? >>>>> >>>> >>>> We're completing https://issues.apache.org/jira/browse/KAFKA-7149 which >>>> should largely reduce the message size (will update the wiki >> accordingly as >>>> well). >>>> >>>> >>>>> >>>>> Boyang >>>>> >>>>> >>>>> From: Guozhang Wang >>>>> Sent: Monday, April 15, 2019 9:20 AM >>>>> To: dev >>>>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams >>>>> >>>>> Hello Jason, >>>>> >>>>> I agree with you th
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Matthias, I'm proposing to change this behavior holistically inside ConsumerCoordinator actually. In other words I'm trying to piggy-back this behavioral fix of KAFKA-4600 along with this KIP, and the motivation for me to do this piggy-backing is that, with incremental rebalancing, there would be partial affected partitions as we are not revoking every body any more. Guozhang On Thu, May 9, 2019 at 6:21 AM Matthias J. Sax wrote: > Thanks Guozhang! > > The simplified upgrade path is great! > > > Just a clarification question about the "Rebalance Callback Error > Handling" -- does this change affect the `ConsumerCoordinator` only if > incremental rebalancing is use? Or does the behavior also change for the > eager rebalancing case? > > > -Matthias > > > On 5/9/19 3:37 AM, Guozhang Wang wrote: > > Hello all, > > > > Thanks for everyone who've shared their feedbacks for this KIP! If > there's > > no further comments I'll start the voting thread by end of tomorrow. > > > > > > Guozhang. > > > > On Wed, May 8, 2019 at 6:36 PM Guozhang Wang wrote: > > > >> Hello Boyang, > >> > >> On Wed, May 1, 2019 at 4:51 PM Boyang Chen wrote: > >> > >>> Hey Guozhang, > >>> > >>> thank you for the great write up. Overall the motivation and changes > >>> LGTM, just some minor comments: > >>> > >>> > >>> 1. In "Consumer Coordinator Algorithm", we could reorder alphabet > >>> points for 3d~3f from ["ready-to-migrate-partitions", > >>> "unknown-but-owned-partitions", "maybe-revoking-partitions"] to > >>> ["maybe-revoking-partitions", "ready-to-migrate-partitions", > >>> "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. > >>> > >> > >> Ack. Updated. > >> > >> > >>> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all > >>> partition upon heartbeat/commit fail. What's the gain here? Do we want > to > >>> keep all partitions running at this moment, to be optimistic for the > case > >>> when no partitions get reassigned? > >>> > >> > >> That's a good catch. When REBALANCE_IN_PROGRESS is received, we can just > >> re-join the group with all the currently owned partitions encoded. > Updated. > >> > >> > >>> 3. In "Recommended Upgrade Procedure", remove extra 'those': " The > >>> 'sticky' assignor works even those there are " > >>> > >> > >> Ack, should be `even when`. > >> > >> > >>> 4. Put two "looking into the future" into a separate category from > >>> migration session. It seems inconsistent for readers to see this > before we > >>> finished discussion for everything. > >>> > >> > >> Ack. > >> > >> > >>> 5. Have we discussed the concern on the serialization? Could the new > >>> metadata we are adding grow larger than the message size cap? > >>> > >> > >> We're completing https://issues.apache.org/jira/browse/KAFKA-7149 which > >> should largely reduce the message size (will update the wiki > accordingly as > >> well). > >> > >> > >>> > >>> Boyang > >>> > >>> > >>> From: Guozhang Wang > >>> Sent: Monday, April 15, 2019 9:20 AM > >>> To: dev > >>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > >>> > >>> Hello Jason, > >>> > >>> I agree with you that for range / round-robin it makes less sense to be > >>> compatible with cooperative rebalance protocol. > >>> > >>> As for StickyAssignor, however, I think it would still be possible to > make > >>> the current implementation to be compatible with cooperative > rebalance. So > >>> after pondering on different options at hand I'm now proposing this > >>> approach as listed in the upgrade section: > >>> > >>> > >>> > https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath > >>> > >>> The idea is to let assignors specify which protocols it would work > with, > >
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Thanks Guozhang! The simplified upgrade path is great! Just a clarification question about the "Rebalance Callback Error Handling" -- does this change affect the `ConsumerCoordinator` only if incremental rebalancing is use? Or does the behavior also change for the eager rebalancing case? -Matthias On 5/9/19 3:37 AM, Guozhang Wang wrote: > Hello all, > > Thanks for everyone who've shared their feedbacks for this KIP! If there's > no further comments I'll start the voting thread by end of tomorrow. > > > Guozhang. > > On Wed, May 8, 2019 at 6:36 PM Guozhang Wang wrote: > >> Hello Boyang, >> >> On Wed, May 1, 2019 at 4:51 PM Boyang Chen wrote: >> >>> Hey Guozhang, >>> >>> thank you for the great write up. Overall the motivation and changes >>> LGTM, just some minor comments: >>> >>> >>> 1. In "Consumer Coordinator Algorithm", we could reorder alphabet >>> points for 3d~3f from ["ready-to-migrate-partitions", >>> "unknown-but-owned-partitions", "maybe-revoking-partitions"] to >>> ["maybe-revoking-partitions", "ready-to-migrate-partitions", >>> "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. >>> >> >> Ack. Updated. >> >> >>> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all >>> partition upon heartbeat/commit fail. What's the gain here? Do we want to >>> keep all partitions running at this moment, to be optimistic for the case >>> when no partitions get reassigned? >>> >> >> That's a good catch. When REBALANCE_IN_PROGRESS is received, we can just >> re-join the group with all the currently owned partitions encoded. Updated. >> >> >>> 3. In "Recommended Upgrade Procedure", remove extra 'those': " The >>> 'sticky' assignor works even those there are " >>> >> >> Ack, should be `even when`. >> >> >>> 4. Put two "looking into the future" into a separate category from >>> migration session. It seems inconsistent for readers to see this before we >>> finished discussion for everything. >>> >> >> Ack. >> >> >>> 5. Have we discussed the concern on the serialization? Could the new >>> metadata we are adding grow larger than the message size cap? >>> >> >> We're completing https://issues.apache.org/jira/browse/KAFKA-7149 which >> should largely reduce the message size (will update the wiki accordingly as >> well). >> >> >>> >>> Boyang >>> >>> >>> From: Guozhang Wang >>> Sent: Monday, April 15, 2019 9:20 AM >>> To: dev >>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams >>> >>> Hello Jason, >>> >>> I agree with you that for range / round-robin it makes less sense to be >>> compatible with cooperative rebalance protocol. >>> >>> As for StickyAssignor, however, I think it would still be possible to make >>> the current implementation to be compatible with cooperative rebalance. So >>> after pondering on different options at hand I'm now proposing this >>> approach as listed in the upgrade section: >>> >>> >>> https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath >>> >>> The idea is to let assignors specify which protocols it would work with, >>> associating with a different name; then the upgrade path would involve a >>> "compatible" protocol which actually still use eager behavior while >>> encoding two assignors if possible. In "Rejected Section" (just to clarify >>> I'm not finalizing it as rejected, just putting it there for now, and if >>> we >>> like this one instead we can always switch them) I listed the other >>> approach we once discussed about, and arguing its cons of duplicated class >>> seems overwhelm the pros of saving the "rebalance.protocol" config. >>> >>> Let me know WDYT. >>> >>> Guozhang >>> >>> On Fri, Apr 12, 2019 at 6:08 PM Jason Gustafson >>> wrote: >>> >>>> Hi Guozhang, >>>> >>>> Responses below: >>>> >>>> 2. The interface's default implementation will just be >&g
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello all, Thanks for everyone who've shared their feedbacks for this KIP! If there's no further comments I'll start the voting thread by end of tomorrow. Guozhang. On Wed, May 8, 2019 at 6:36 PM Guozhang Wang wrote: > Hello Boyang, > > On Wed, May 1, 2019 at 4:51 PM Boyang Chen wrote: > >> Hey Guozhang, >> >> thank you for the great write up. Overall the motivation and changes >> LGTM, just some minor comments: >> >> >> 1. In "Consumer Coordinator Algorithm", we could reorder alphabet >> points for 3d~3f from ["ready-to-migrate-partitions", >> "unknown-but-owned-partitions", "maybe-revoking-partitions"] to >> ["maybe-revoking-partitions", "ready-to-migrate-partitions", >> "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. >> > > Ack. Updated. > > >> 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all >> partition upon heartbeat/commit fail. What's the gain here? Do we want to >> keep all partitions running at this moment, to be optimistic for the case >> when no partitions get reassigned? >> > > That's a good catch. When REBALANCE_IN_PROGRESS is received, we can just > re-join the group with all the currently owned partitions encoded. Updated. > > >> 3. In "Recommended Upgrade Procedure", remove extra 'those': " The >> 'sticky' assignor works even those there are " >> > > Ack, should be `even when`. > > >> 4. Put two "looking into the future" into a separate category from >> migration session. It seems inconsistent for readers to see this before we >> finished discussion for everything. >> > > Ack. > > >> 5. Have we discussed the concern on the serialization? Could the new >> metadata we are adding grow larger than the message size cap? >> > > We're completing https://issues.apache.org/jira/browse/KAFKA-7149 which > should largely reduce the message size (will update the wiki accordingly as > well). > > >> >> Boyang >> >> >> From: Guozhang Wang >> Sent: Monday, April 15, 2019 9:20 AM >> To: dev >> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams >> >> Hello Jason, >> >> I agree with you that for range / round-robin it makes less sense to be >> compatible with cooperative rebalance protocol. >> >> As for StickyAssignor, however, I think it would still be possible to make >> the current implementation to be compatible with cooperative rebalance. So >> after pondering on different options at hand I'm now proposing this >> approach as listed in the upgrade section: >> >> >> https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath >> >> The idea is to let assignors specify which protocols it would work with, >> associating with a different name; then the upgrade path would involve a >> "compatible" protocol which actually still use eager behavior while >> encoding two assignors if possible. In "Rejected Section" (just to clarify >> I'm not finalizing it as rejected, just putting it there for now, and if >> we >> like this one instead we can always switch them) I listed the other >> approach we once discussed about, and arguing its cons of duplicated class >> seems overwhelm the pros of saving the "rebalance.protocol" config. >> >> Let me know WDYT. >> >> Guozhang >> >> On Fri, Apr 12, 2019 at 6:08 PM Jason Gustafson >> wrote: >> >> > Hi Guozhang, >> > >> > Responses below: >> > >> > 2. The interface's default implementation will just be >> > > `onPartitionRevoked`, so for user's instantiation if they do not make >> any >> > > code changes they should be able to recompile the code and continue. >> > >> > >> > Ack, makes sense. >> > >> > 4. Hmm.. not sure if it will work. The main issue is that the >> > > consumer-coordinator behavior (whether to revoke all or none at >> > > onRebalancePrepare) is independent of the selected protocol's assignor >> > > (eager or cooperative), so even if the assignor is selected to be the >> > > old-versioned one, we will still not revoke at the >> consumer-coordinator >> > > layer and hence has the same risk of migrating still-owned partitions, >&g
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Boyang, On Wed, May 1, 2019 at 4:51 PM Boyang Chen wrote: > Hey Guozhang, > > thank you for the great write up. Overall the motivation and changes LGTM, > just some minor comments: > > > 1. In "Consumer Coordinator Algorithm", we could reorder alphabet > points for 3d~3f from ["ready-to-migrate-partitions", > "unknown-but-owned-partitions", "maybe-revoking-partitions"] to > ["maybe-revoking-partitions", "ready-to-migrate-partitions", > "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. > Ack. Updated. > 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all > partition upon heartbeat/commit fail. What's the gain here? Do we want to > keep all partitions running at this moment, to be optimistic for the case > when no partitions get reassigned? > That's a good catch. When REBALANCE_IN_PROGRESS is received, we can just re-join the group with all the currently owned partitions encoded. Updated. > 3. In "Recommended Upgrade Procedure", remove extra 'those': " The > 'sticky' assignor works even those there are " > Ack, should be `even when`. > 4. Put two "looking into the future" into a separate category from > migration session. It seems inconsistent for readers to see this before we > finished discussion for everything. > Ack. > 5. Have we discussed the concern on the serialization? Could the new > metadata we are adding grow larger than the message size cap? > We're completing https://issues.apache.org/jira/browse/KAFKA-7149 which should largely reduce the message size (will update the wiki accordingly as well). > > Boyang > > > From: Guozhang Wang > Sent: Monday, April 15, 2019 9:20 AM > To: dev > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > Hello Jason, > > I agree with you that for range / round-robin it makes less sense to be > compatible with cooperative rebalance protocol. > > As for StickyAssignor, however, I think it would still be possible to make > the current implementation to be compatible with cooperative rebalance. So > after pondering on different options at hand I'm now proposing this > approach as listed in the upgrade section: > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath > > The idea is to let assignors specify which protocols it would work with, > associating with a different name; then the upgrade path would involve a > "compatible" protocol which actually still use eager behavior while > encoding two assignors if possible. In "Rejected Section" (just to clarify > I'm not finalizing it as rejected, just putting it there for now, and if we > like this one instead we can always switch them) I listed the other > approach we once discussed about, and arguing its cons of duplicated class > seems overwhelm the pros of saving the "rebalance.protocol" config. > > Let me know WDYT. > > Guozhang > > On Fri, Apr 12, 2019 at 6:08 PM Jason Gustafson > wrote: > > > Hi Guozhang, > > > > Responses below: > > > > 2. The interface's default implementation will just be > > > `onPartitionRevoked`, so for user's instantiation if they do not make > any > > > code changes they should be able to recompile the code and continue. > > > > > > Ack, makes sense. > > > > 4. Hmm.. not sure if it will work. The main issue is that the > > > consumer-coordinator behavior (whether to revoke all or none at > > > onRebalancePrepare) is independent of the selected protocol's assignor > > > (eager or cooperative), so even if the assignor is selected to be the > > > old-versioned one, we will still not revoke at the consumer-coordinator > > > layer and hence has the same risk of migrating still-owned partitions, > > > right? > > > > > > Yeah, basically we would have to push the eager/cooperative logic into > the > > PartitionAssignor itself and make the consumer aware of the rebalance > > protocol it is compatible with. As long as an eager protocol _could_ be > > selected, the consumer would have to be pessimistic and do eager > > revocation. But if all the assignors configured in the consumer support > > cooperative reassignment, then either 1) a cooperative protocol will be > > selected and cooperative revocation can be safely used, or 2) if the rest > > of the group does not support it, then the consumer will simply fai
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hey Guozhang, thank you for the great write up. Overall the motivation and changes LGTM, just some minor comments: 1. In "Consumer Coordinator Algorithm", we could reorder alphabet points for 3d~3f from ["ready-to-migrate-partitions", "unknown-but-owned-partitions", "maybe-revoking-partitions"] to ["maybe-revoking-partitions", "ready-to-migrate-partitions", "unknown-but-owned-partitions"] in order to be consistent with 3c1~3. 2. In "Consumer Coordinator Algorithm", 1c suggests to revoke all partition upon heartbeat/commit fail. What's the gain here? Do we want to keep all partitions running at this moment, to be optimistic for the case when no partitions get reassigned? 3. In "Recommended Upgrade Procedure", remove extra 'those': " The 'sticky' assignor works even those there are " 4. Put two "looking into the future" into a separate category from migration session. It seems inconsistent for readers to see this before we finished discussion for everything. 5. Have we discussed the concern on the serialization? Could the new metadata we are adding grow larger than the message size cap? Boyang ________________________ From: Guozhang Wang Sent: Monday, April 15, 2019 9:20 AM To: dev Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams Hello Jason, I agree with you that for range / round-robin it makes less sense to be compatible with cooperative rebalance protocol. As for StickyAssignor, however, I think it would still be possible to make the current implementation to be compatible with cooperative rebalance. So after pondering on different options at hand I'm now proposing this approach as listed in the upgrade section: https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol#KIP-429:KafkaConsumerIncrementalRebalanceProtocol-CompatibilityandUpgradePath The idea is to let assignors specify which protocols it would work with, associating with a different name; then the upgrade path would involve a "compatible" protocol which actually still use eager behavior while encoding two assignors if possible. In "Rejected Section" (just to clarify I'm not finalizing it as rejected, just putting it there for now, and if we like this one instead we can always switch them) I listed the other approach we once discussed about, and arguing its cons of duplicated class seems overwhelm the pros of saving the "rebalance.protocol" config. Let me know WDYT. Guozhang On Fri, Apr 12, 2019 at 6:08 PM Jason Gustafson wrote: > Hi Guozhang, > > Responses below: > > 2. The interface's default implementation will just be > > `onPartitionRevoked`, so for user's instantiation if they do not make any > > code changes they should be able to recompile the code and continue. > > > Ack, makes sense. > > 4. Hmm.. not sure if it will work. The main issue is that the > > consumer-coordinator behavior (whether to revoke all or none at > > onRebalancePrepare) is independent of the selected protocol's assignor > > (eager or cooperative), so even if the assignor is selected to be the > > old-versioned one, we will still not revoke at the consumer-coordinator > > layer and hence has the same risk of migrating still-owned partitions, > > right? > > > Yeah, basically we would have to push the eager/cooperative logic into the > PartitionAssignor itself and make the consumer aware of the rebalance > protocol it is compatible with. As long as an eager protocol _could_ be > selected, the consumer would have to be pessimistic and do eager > revocation. But if all the assignors configured in the consumer support > cooperative reassignment, then either 1) a cooperative protocol will be > selected and cooperative revocation can be safely used, or 2) if the rest > of the group does not support it, then the consumer will simply fail. > > Another point which you raised offline and I will repeat here is that this > proposal's benefit is mostly limited to sticky assignment logic. Arguably > the range assignor may have some incidental stickiness, particularly if the > group is rebalancing for a newly created or deleted topic. For other cases, > the proposal is mostly additional overhead since it takes an additional > rebalance and many of the partitions will move. Perhaps it doesn't make as > much sense to use the cooperative protocol for strategies like range and > round-robin. That kind of argues in favor of pushing some of the control > into the assignor itself. Maybe we would not bother creating > CooperativeRange as I suggested above, but it would make sense to create a > cooperative version of the sticky assignment strategy. I thought we might > have to create a
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
re some drawbacks that can not be > > addressed by rebalance delay still, hence still voted KIP-345 (some more > > details can be found on the discussion thread of KIP-345 itself). One > > example is that as the instance resumes, its member id will be empty so > we > > are still relying on assignor to give it the assignment from the old > > member-id while keeping all other member's assignment unchanged. > > > > 40). Incomplete sentence, I've updated it. > > > > 50). Here's my idea: suppose we augment the join group schema with > > `protocol version` in 2.3, and then with both brokers and clients being > in > > version 2.3+, on the first rolling bounce where subscription and > assignment > > schema and / or user metadata has changed, this protocol version will be > > bumped. On the broker side, when receiving all member's join-group > request, > > it will choose the one that has the highest protocol version (also it > > assumes higher versioned protocol is always backward compatible, i.e. the > > coordinator can recognize lower versioned protocol as well) and select it > > as the leader. Then the leader can decide, based on its received and > > deserialized subscription information, how to assign partitions and how > to > > encode the assignment accordingly so that everyone can understand it. > With > > this, in Streams for example, no version probing would be needed since we > > are guaranteed the leader knows everyone's version -- again it is > assuming > > that higher versioned protocol is always backward compatible -- and hence > > can successfully do the assignment at that round. > > > > 60). My bad, this section was not updated while the design was evolved, > > I've updated it. > > > > > > On Tue, Apr 9, 2019 at 7:22 PM Boyang Chen wrote: > > > > > > > > Thanks for the review Matthias! My 2-cent on the rebalance delay is > that > > > it is a rather fixed trade-off between > > > > > > task availability and resource shuffling. If we eventually trigger > > > rebalance after rolling bounce, certain consumer > > > > > > setup is still faced with global shuffles, for example member.id > ranking > > > based round robin strategy, as rejoining dynamic > > > > > > members will be assigned with new member.id which reorders the > > > assignment. So I think the primary goal of incremental > > > > > > rebalancing is still improving the cluster availability during > rebalance, > > > because it didn't revoke any partition during this > > > > > > process. Also, the perk is minimum configuration requirement :) > > > > > > > > > Best, > > > > > > Boyang > > > > > > > > > From: Matthias J. Sax > > > Sent: Tuesday, April 9, 2019 7:47 AM > > > To: dev > > > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > > > > > Thank for the KIP, Boyang and Guozhang! > > > > > > > > > I made an initial pass and have some questions/comments. One high level > > > comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > > > use case a little bit (at least in the presentation). It might be > > > helpful to separate both cases clearly, or maybe limit the scope to > > > plain consumer only. > > > > > > > > > > > > 10) For `PartitionAssignor.Assignment`: It seems we need a new method > > > `List revokedPartitions()` ? > > > > > > > > > > > > 20) In Section "Consumer Coordinator Algorithm" > > > > > > Bullet point "1a)": If the subscription changes and a topic is > > > removed from the subscription, why do we not revoke the partitions? > > > > > > Bullet point "1a)": What happens is a topic is deleted (or a > > > partition is removed/deleted from a topic)? Should we call the new > > > `onPartitionsEmigrated()` callback for this case? > > > > > > Bullet point "2b)" Should we update the `PartitionAssignor` > > > interface to pass in the "old assignment" as third parameter into > > > `assign()`? > > > > > > > > > > > > 30) Rebalance delay (as used in KIP-415): Could a rebalance delay > > > subsume KIP-345? Configuring static members is rather complicated, and > I > > > am wondering if a rebalance delay would be sufficient? > > >
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
gt; >> `protocol version` in 2.3, and then with both brokers and clients being > in > >> version 2.3+, on the first rolling bounce where subscription and > assignment > >> schema and / or user metadata has changed, this protocol version will be > >> bumped. On the broker side, when receiving all member's join-group > request, > >> it will choose the one that has the highest protocol version (also it > >> assumes higher versioned protocol is always backward compatible, i.e. > the > >> coordinator can recognize lower versioned protocol as well) and select > it > >> as the leader. Then the leader can decide, based on its received and > >> deserialized subscription information, how to assign partitions and how > to > >> encode the assignment accordingly so that everyone can understand it. > With > >> this, in Streams for example, no version probing would be needed since > we > >> are guaranteed the leader knows everyone's version -- again it is > assuming > >> that higher versioned protocol is always backward compatible -- and > hence > >> can successfully do the assignment at that round. > >> > >> 60). My bad, this section was not updated while the design was evolved, > >> I've updated it. > >> > >> > >> On Tue, Apr 9, 2019 at 7:22 PM Boyang Chen wrote: > >> > >>> > >>> Thanks for the review Matthias! My 2-cent on the rebalance delay is > that > >>> it is a rather fixed trade-off between > >>> > >>> task availability and resource shuffling. If we eventually trigger > >>> rebalance after rolling bounce, certain consumer > >>> > >>> setup is still faced with global shuffles, for example member.id > ranking > >>> based round robin strategy, as rejoining dynamic > >>> > >>> members will be assigned with new member.id which reorders the > >>> assignment. So I think the primary goal of incremental > >>> > >>> rebalancing is still improving the cluster availability during > rebalance, > >>> because it didn't revoke any partition during this > >>> > >>> process. Also, the perk is minimum configuration requirement :) > >>> > >>> > >>> Best, > >>> > >>> Boyang > >>> > >>> > >>> From: Matthias J. Sax > >>> Sent: Tuesday, April 9, 2019 7:47 AM > >>> To: dev > >>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > >>> > >>> Thank for the KIP, Boyang and Guozhang! > >>> > >>> > >>> I made an initial pass and have some questions/comments. One high level > >>> comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > >>> use case a little bit (at least in the presentation). It might be > >>> helpful to separate both cases clearly, or maybe limit the scope to > >>> plain consumer only. > >>> > >>> > >>> > >>> 10) For `PartitionAssignor.Assignment`: It seems we need a new method > >>> `List revokedPartitions()` ? > >>> > >>> > >>> > >>> 20) In Section "Consumer Coordinator Algorithm" > >>> > >>> Bullet point "1a)": If the subscription changes and a topic is > >>> removed from the subscription, why do we not revoke the partitions? > >>> > >>> Bullet point "1a)": What happens is a topic is deleted (or a > >>> partition is removed/deleted from a topic)? Should we call the new > >>> `onPartitionsEmigrated()` callback for this case? > >>> > >>> Bullet point "2b)" Should we update the `PartitionAssignor` > >>> interface to pass in the "old assignment" as third parameter into > >>> `assign()`? > >>> > >>> > >>> > >>> 30) Rebalance delay (as used in KIP-415): Could a rebalance delay > >>> subsume KIP-345? Configuring static members is rather complicated, and > I > >>> am wondering if a rebalance delay would be sufficient? > >>> > >>> > >>> > >>> 40) Quote: "otherwise the we would fall into the case 3.b) forever." > >>> > >>> What is "case 3.b" ? > >>> > >>> > >>> > >>&
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
ybe we would not bother creating > CooperativeRange as I suggested above, but it would make sense to create a > cooperative version of the sticky assignment strategy. I thought we might > have to create a new sticky assignor anyway because I can't see how we > would get compatible behavior mixing with the old version anyway. > > Thanks, > Jason > > > On Thu, Apr 11, 2019 at 5:53 PM Guozhang Wang wrote: > >> Hello Matthias: >> >> Thanks for your review. >> >> The background section uses streams assignor as well as the consumer's own >> stick assignor as examples illustrating the situation, but this KIP is for >> consumer coordinator itself, and the rest of the paragraph did not talk >> about Streams any more. If you feel it's a bit distracted I can remove >> those examples. >> >> 10). While working on the PR I realized that the revoked partitions on >> assignment is not needed (this is being discussed on the PR itself: >> https://github.com/apache/kafka/pull/6528#issuecomment-480009890 >> >> 20). 1.a. Good question, I've updated the wiki to let the consumer's >> cleanup assignment and re-join, and not letting assignor making any >> proactive changes. The idea is to keep logic simpler and not doing any >> "split brain" stuff. >> >> 20). 2.b. No we do not need, since the owned-partitions will be part of the >> Subscription passed in to assign() already. >> >> 30). As Boyang mentioned, there are some drawbacks that can not be >> addressed by rebalance delay still, hence still voted KIP-345 (some more >> details can be found on the discussion thread of KIP-345 itself). One >> example is that as the instance resumes, its member id will be empty so we >> are still relying on assignor to give it the assignment from the old >> member-id while keeping all other member's assignment unchanged. >> >> 40). Incomplete sentence, I've updated it. >> >> 50). Here's my idea: suppose we augment the join group schema with >> `protocol version` in 2.3, and then with both brokers and clients being in >> version 2.3+, on the first rolling bounce where subscription and assignment >> schema and / or user metadata has changed, this protocol version will be >> bumped. On the broker side, when receiving all member's join-group request, >> it will choose the one that has the highest protocol version (also it >> assumes higher versioned protocol is always backward compatible, i.e. the >> coordinator can recognize lower versioned protocol as well) and select it >> as the leader. Then the leader can decide, based on its received and >> deserialized subscription information, how to assign partitions and how to >> encode the assignment accordingly so that everyone can understand it. With >> this, in Streams for example, no version probing would be needed since we >> are guaranteed the leader knows everyone's version -- again it is assuming >> that higher versioned protocol is always backward compatible -- and hence >> can successfully do the assignment at that round. >> >> 60). My bad, this section was not updated while the design was evolved, >> I've updated it. >> >> >> On Tue, Apr 9, 2019 at 7:22 PM Boyang Chen wrote: >> >>> >>> Thanks for the review Matthias! My 2-cent on the rebalance delay is that >>> it is a rather fixed trade-off between >>> >>> task availability and resource shuffling. If we eventually trigger >>> rebalance after rolling bounce, certain consumer >>> >>> setup is still faced with global shuffles, for example member.id ranking >>> based round robin strategy, as rejoining dynamic >>> >>> members will be assigned with new member.id which reorders the >>> assignment. So I think the primary goal of incremental >>> >>> rebalancing is still improving the cluster availability during rebalance, >>> because it didn't revoke any partition during this >>> >>> process. Also, the perk is minimum configuration requirement :) >>> >>> >>> Best, >>> >>> Boyang >>> >>> >>> From: Matthias J. Sax >>> Sent: Tuesday, April 9, 2019 7:47 AM >>> To: dev >>> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams >>> >>> Thank for the KIP, Boyang and Guozhang! >>> >>> >>> I made an initial pass and have some questions/comments. One high level >>> comment: it seems that the KIP "mixes" plain consumer and Kafka Streams >>&
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
d, this section was not updated while the design was evolved, > I've updated it. > > > On Tue, Apr 9, 2019 at 7:22 PM Boyang Chen wrote: > > > > > Thanks for the review Matthias! My 2-cent on the rebalance delay is that > > it is a rather fixed trade-off between > > > > task availability and resource shuffling. If we eventually trigger > > rebalance after rolling bounce, certain consumer > > > > setup is still faced with global shuffles, for example member.id ranking > > based round robin strategy, as rejoining dynamic > > > > members will be assigned with new member.id which reorders the > > assignment. So I think the primary goal of incremental > > > > rebalancing is still improving the cluster availability during rebalance, > > because it didn't revoke any partition during this > > > > process. Also, the perk is minimum configuration requirement :) > > > > > > Best, > > > > Boyang > > > > > > From: Matthias J. Sax > > Sent: Tuesday, April 9, 2019 7:47 AM > > To: dev > > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > > > Thank for the KIP, Boyang and Guozhang! > > > > > > I made an initial pass and have some questions/comments. One high level > > comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > > use case a little bit (at least in the presentation). It might be > > helpful to separate both cases clearly, or maybe limit the scope to > > plain consumer only. > > > > > > > > 10) For `PartitionAssignor.Assignment`: It seems we need a new method > > `List revokedPartitions()` ? > > > > > > > > 20) In Section "Consumer Coordinator Algorithm" > > > > Bullet point "1a)": If the subscription changes and a topic is > > removed from the subscription, why do we not revoke the partitions? > > > > Bullet point "1a)": What happens is a topic is deleted (or a > > partition is removed/deleted from a topic)? Should we call the new > > `onPartitionsEmigrated()` callback for this case? > > > > Bullet point "2b)" Should we update the `PartitionAssignor` > > interface to pass in the "old assignment" as third parameter into > > `assign()`? > > > > > > > > 30) Rebalance delay (as used in KIP-415): Could a rebalance delay > > subsume KIP-345? Configuring static members is rather complicated, and I > > am wondering if a rebalance delay would be sufficient? > > > > > > > > 40) Quote: "otherwise the we would fall into the case 3.b) forever." > > > > What is "case 3.b" ? > > > > > > > > 50) Section "Looking into the Future" > > > > Nit: the new "ProtocolVersion" field is missing in the first line > > describing "JoinGroupRequest" > > > > > This can also help saving "version probing" cost on Streams as well. > > > > How does this relate to Kafka Streams "version probing" implementation? > > How can we exploit the new `ProtocolVersion` in Streams to improve > > "version probing" ? I have a rough idea, but would like to hear more > > details. > > > > > > > > 60) Section "Recommended Upgrade Procedure" > > > > > Set the `stream.rebalancing.mode` to `upgrading`, which will force the > > stream application to stay with protocol type "consumer". > > > > This config is not discussed in the KIP and appears in this section > > without context. Can you elaborate about it? > > > > > > > > -Matthias > > > > > > > > > > On 3/29/19 6:20 PM, Guozhang Wang wrote: > > > Bump up on this discussion thread. I've added a few new drawings for > > better > > > illustration, would really appreciate your feedbacks. > > > > > > > > > Guozhang > > > > > > On Wed, Mar 20, 2019 at 6:17 PM Guozhang Wang > > wrote: > > > > > >> Hello Boyang, > > >> > > >> I've made another thorough pass over this KIP and I'd like to spilt it > > >> into two parts: the first part, covered in KIP-429 would be touching > on > > >> Consumer Coordinator only to have incremental rebalance protocol in > > place. > > >> The second part (for now I've reserved KIP number 444 for it) would > > contain > > >> all the changes on StreamsPart
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Matthias: Thanks for your review. The background section uses streams assignor as well as the consumer's own stick assignor as examples illustrating the situation, but this KIP is for consumer coordinator itself, and the rest of the paragraph did not talk about Streams any more. If you feel it's a bit distracted I can remove those examples. 10). While working on the PR I realized that the revoked partitions on assignment is not needed (this is being discussed on the PR itself: https://github.com/apache/kafka/pull/6528#issuecomment-480009890 20). 1.a. Good question, I've updated the wiki to let the consumer's cleanup assignment and re-join, and not letting assignor making any proactive changes. The idea is to keep logic simpler and not doing any "split brain" stuff. 20). 2.b. No we do not need, since the owned-partitions will be part of the Subscription passed in to assign() already. 30). As Boyang mentioned, there are some drawbacks that can not be addressed by rebalance delay still, hence still voted KIP-345 (some more details can be found on the discussion thread of KIP-345 itself). One example is that as the instance resumes, its member id will be empty so we are still relying on assignor to give it the assignment from the old member-id while keeping all other member's assignment unchanged. 40). Incomplete sentence, I've updated it. 50). Here's my idea: suppose we augment the join group schema with `protocol version` in 2.3, and then with both brokers and clients being in version 2.3+, on the first rolling bounce where subscription and assignment schema and / or user metadata has changed, this protocol version will be bumped. On the broker side, when receiving all member's join-group request, it will choose the one that has the highest protocol version (also it assumes higher versioned protocol is always backward compatible, i.e. the coordinator can recognize lower versioned protocol as well) and select it as the leader. Then the leader can decide, based on its received and deserialized subscription information, how to assign partitions and how to encode the assignment accordingly so that everyone can understand it. With this, in Streams for example, no version probing would be needed since we are guaranteed the leader knows everyone's version -- again it is assuming that higher versioned protocol is always backward compatible -- and hence can successfully do the assignment at that round. 60). My bad, this section was not updated while the design was evolved, I've updated it. On Tue, Apr 9, 2019 at 7:22 PM Boyang Chen wrote: > > Thanks for the review Matthias! My 2-cent on the rebalance delay is that > it is a rather fixed trade-off between > > task availability and resource shuffling. If we eventually trigger > rebalance after rolling bounce, certain consumer > > setup is still faced with global shuffles, for example member.id ranking > based round robin strategy, as rejoining dynamic > > members will be assigned with new member.id which reorders the > assignment. So I think the primary goal of incremental > > rebalancing is still improving the cluster availability during rebalance, > because it didn't revoke any partition during this > > process. Also, the perk is minimum configuration requirement :) > > > Best, > > Boyang > > > From: Matthias J. Sax > Sent: Tuesday, April 9, 2019 7:47 AM > To: dev > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > Thank for the KIP, Boyang and Guozhang! > > > I made an initial pass and have some questions/comments. One high level > comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > use case a little bit (at least in the presentation). It might be > helpful to separate both cases clearly, or maybe limit the scope to > plain consumer only. > > > > 10) For `PartitionAssignor.Assignment`: It seems we need a new method > `List revokedPartitions()` ? > > > > 20) In Section "Consumer Coordinator Algorithm" > > Bullet point "1a)": If the subscription changes and a topic is > removed from the subscription, why do we not revoke the partitions? > > Bullet point "1a)": What happens is a topic is deleted (or a > partition is removed/deleted from a topic)? Should we call the new > `onPartitionsEmigrated()` callback for this case? > > Bullet point "2b)" Should we update the `PartitionAssignor` > interface to pass in the "old assignment" as third parameter into > `assign()`? > > > > 30) Rebalance delay (as used in KIP-415): Could a rebalance delay > subsume KIP-345? Configuring static members is rather complicated, and I > am wondering if a rebalance delay would be sufficient? > > > > 40) Quote: "ot
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Jason, 1. Yeah I think `emigrate` is not a very good term maybe. On the other hand `evicted` is quite similar to `revoked` to me.. I'd like to have another proposal of `onPartitionLost` actually. 2. The interface's default implementation will just be `onPartitionRevoked`, so for user's instantiation if they do not make any code changes they should be able to recompile the code and continue. 3. Yes: this is to reduce the rebalance notification latency (green bar in the digram). 4. Hmm.. not sure if it will work. The main issue is that the consumer-coordinator behavior (whether to revoke all or none at onRebalancePrepare) is independent of the selected protocol's assignor (eager or cooperative), so even if the assignor is selected to be the old-versioned one, we will still not revoke at the consumer-coordinator layer and hence has the same risk of migrating still-owned partitions, right? 5. Yup good point, I will add it to the PartitionAssignor interface. Guozhang On Wed, Apr 10, 2019 at 4:03 PM Jason Gustafson wrote: > Hi Guozhang and Boyang, > > Thanks for the KIP. A few comments/questions below: > > 1. More of a nitpick, but `onPartitionsEmigrated` is not a very clear name. > How about `onPartitionsEvicted`? Or even perhaps `onMembershipLost`? > 2. For `onPartitionsEmigrated`, how will we maintain compatibility with the > old behavior? Seems like we might need an extension of > `ConsumerRebalanceListener` with the new method. Otherwise we won't know if > the application is expecting the old behavior. > 3. Just making sure I understand this, but the reason we need the error > code in the assignment is that the revoked partitions might be empty for > some members and non-empty for others. We want all members to rejoin > quickly even if they have no revoked partitions. Is that right? > 4. I wanted to suggest an alternative approach for dealing with > compatibility and the upgrade problem. In fact, the consumer already has a > mechanism to change the assignment logic. Users can provide multiple > PartitionAssignor implementations in the `partition.assignment.strategy` > configuration. The coordinator will only select one which is supported by > all members of the group. Rather than adding the new `rebalance.protocol` > config, could we not reuse this mechanism? To support this, we would > basically create new assignor implementations. For example, > CooperativeRoundRobin instead of the usual RoundRobin. I think the benefit > is that it is quite a bit easier to reason about the upgrade state when not > all consumers have been updated. We are guaranteed that all members are > following the same logic. My feeling is that this will be a less error > prone solution since it depends less on state outside the system (i.e. the > respective `rebalance.protocol` configurations for all members in the group > and binary compatibility). The downside is that it will take more effort > for PartitionAssignor implementations to get a benefit from this improved > logic. But it's really hard to say that the new assignment logic would be > compatible with a custom assignor in any case. > 5. Where does the new ProtocolVersion come from in the new JoinGroup > schema? I guess we need a new API on the PartitionAssignor interface? > > Thanks, > Jason > > > On Mon, Apr 8, 2019 at 9:39 PM Boyang Chen wrote: > > > > > Thanks for the review Matthias! My 2-cent on the rebalance delay is that > > it is a rather fixed trade-off between > > > > task availability and resource shuffling. If we eventually trigger > > rebalance after rolling bounce, certain consumer > > > > setup is still faced with global shuffles, for example member.id ranking > > based round robin strategy, as rejoining dynamic > > > > members will be assigned with new member.id which reorders the > > assignment. So I think the primary goal of incremental > > > > rebalancing is still improving the cluster availability during rebalance, > > because it didn't revoke any partition during this > > > > process. Also, the perk is minimum configuration requirement :) > > > > > > Best, > > > > Boyang > > > > > > From: Matthias J. Sax > > Sent: Tuesday, April 9, 2019 7:47 AM > > To: dev > > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > > > Thank for the KIP, Boyang and Guozhang! > > > > > > I made an initial pass and have some questions/comments. One high level > > comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > > use case a little bit (at least in the presentation). It might be > > helpful to separate both cases clearly, or may
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hi Guozhang and Boyang, Thanks for the KIP. A few comments/questions below: 1. More of a nitpick, but `onPartitionsEmigrated` is not a very clear name. How about `onPartitionsEvicted`? Or even perhaps `onMembershipLost`? 2. For `onPartitionsEmigrated`, how will we maintain compatibility with the old behavior? Seems like we might need an extension of `ConsumerRebalanceListener` with the new method. Otherwise we won't know if the application is expecting the old behavior. 3. Just making sure I understand this, but the reason we need the error code in the assignment is that the revoked partitions might be empty for some members and non-empty for others. We want all members to rejoin quickly even if they have no revoked partitions. Is that right? 4. I wanted to suggest an alternative approach for dealing with compatibility and the upgrade problem. In fact, the consumer already has a mechanism to change the assignment logic. Users can provide multiple PartitionAssignor implementations in the `partition.assignment.strategy` configuration. The coordinator will only select one which is supported by all members of the group. Rather than adding the new `rebalance.protocol` config, could we not reuse this mechanism? To support this, we would basically create new assignor implementations. For example, CooperativeRoundRobin instead of the usual RoundRobin. I think the benefit is that it is quite a bit easier to reason about the upgrade state when not all consumers have been updated. We are guaranteed that all members are following the same logic. My feeling is that this will be a less error prone solution since it depends less on state outside the system (i.e. the respective `rebalance.protocol` configurations for all members in the group and binary compatibility). The downside is that it will take more effort for PartitionAssignor implementations to get a benefit from this improved logic. But it's really hard to say that the new assignment logic would be compatible with a custom assignor in any case. 5. Where does the new ProtocolVersion come from in the new JoinGroup schema? I guess we need a new API on the PartitionAssignor interface? Thanks, Jason On Mon, Apr 8, 2019 at 9:39 PM Boyang Chen wrote: > > Thanks for the review Matthias! My 2-cent on the rebalance delay is that > it is a rather fixed trade-off between > > task availability and resource shuffling. If we eventually trigger > rebalance after rolling bounce, certain consumer > > setup is still faced with global shuffles, for example member.id ranking > based round robin strategy, as rejoining dynamic > > members will be assigned with new member.id which reorders the > assignment. So I think the primary goal of incremental > > rebalancing is still improving the cluster availability during rebalance, > because it didn't revoke any partition during this > > process. Also, the perk is minimum configuration requirement :) > > > Best, > > Boyang > > > From: Matthias J. Sax > Sent: Tuesday, April 9, 2019 7:47 AM > To: dev > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > Thank for the KIP, Boyang and Guozhang! > > > I made an initial pass and have some questions/comments. One high level > comment: it seems that the KIP "mixes" plain consumer and Kafka Streams > use case a little bit (at least in the presentation). It might be > helpful to separate both cases clearly, or maybe limit the scope to > plain consumer only. > > > > 10) For `PartitionAssignor.Assignment`: It seems we need a new method > `List revokedPartitions()` ? > > > > 20) In Section "Consumer Coordinator Algorithm" > > Bullet point "1a)": If the subscription changes and a topic is > removed from the subscription, why do we not revoke the partitions? > > Bullet point "1a)": What happens is a topic is deleted (or a > partition is removed/deleted from a topic)? Should we call the new > `onPartitionsEmigrated()` callback for this case? > > Bullet point "2b)" Should we update the `PartitionAssignor` > interface to pass in the "old assignment" as third parameter into > `assign()`? > > > > 30) Rebalance delay (as used in KIP-415): Could a rebalance delay > subsume KIP-345? Configuring static members is rather complicated, and I > am wondering if a rebalance delay would be sufficient? > > > > 40) Quote: "otherwise the we would fall into the case 3.b) forever." > > What is "case 3.b" ? > > > > 50) Section "Looking into the Future" > > Nit: the new "ProtocolVersion" field is missing in the first line > describing "JoinGroupRequest" > > > This can also help saving "version probing"
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Thanks for the review Matthias! My 2-cent on the rebalance delay is that it is a rather fixed trade-off between task availability and resource shuffling. If we eventually trigger rebalance after rolling bounce, certain consumer setup is still faced with global shuffles, for example member.id ranking based round robin strategy, as rejoining dynamic members will be assigned with new member.id which reorders the assignment. So I think the primary goal of incremental rebalancing is still improving the cluster availability during rebalance, because it didn't revoke any partition during this process. Also, the perk is minimum configuration requirement :) Best, Boyang From: Matthias J. Sax Sent: Tuesday, April 9, 2019 7:47 AM To: dev Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams Thank for the KIP, Boyang and Guozhang! I made an initial pass and have some questions/comments. One high level comment: it seems that the KIP "mixes" plain consumer and Kafka Streams use case a little bit (at least in the presentation). It might be helpful to separate both cases clearly, or maybe limit the scope to plain consumer only. 10) For `PartitionAssignor.Assignment`: It seems we need a new method `List revokedPartitions()` ? 20) In Section "Consumer Coordinator Algorithm" Bullet point "1a)": If the subscription changes and a topic is removed from the subscription, why do we not revoke the partitions? Bullet point "1a)": What happens is a topic is deleted (or a partition is removed/deleted from a topic)? Should we call the new `onPartitionsEmigrated()` callback for this case? Bullet point "2b)" Should we update the `PartitionAssignor` interface to pass in the "old assignment" as third parameter into `assign()`? 30) Rebalance delay (as used in KIP-415): Could a rebalance delay subsume KIP-345? Configuring static members is rather complicated, and I am wondering if a rebalance delay would be sufficient? 40) Quote: "otherwise the we would fall into the case 3.b) forever." What is "case 3.b" ? 50) Section "Looking into the Future" Nit: the new "ProtocolVersion" field is missing in the first line describing "JoinGroupRequest" > This can also help saving "version probing" cost on Streams as well. How does this relate to Kafka Streams "version probing" implementation? How can we exploit the new `ProtocolVersion` in Streams to improve "version probing" ? I have a rough idea, but would like to hear more details. 60) Section "Recommended Upgrade Procedure" > Set the `stream.rebalancing.mode` to `upgrading`, which will force the stream > application to stay with protocol type "consumer". This config is not discussed in the KIP and appears in this section without context. Can you elaborate about it? -Matthias On 3/29/19 6:20 PM, Guozhang Wang wrote: > Bump up on this discussion thread. I've added a few new drawings for better > illustration, would really appreciate your feedbacks. > > > Guozhang > > On Wed, Mar 20, 2019 at 6:17 PM Guozhang Wang wrote: > >> Hello Boyang, >> >> I've made another thorough pass over this KIP and I'd like to spilt it >> into two parts: the first part, covered in KIP-429 would be touching on >> Consumer Coordinator only to have incremental rebalance protocol in place. >> The second part (for now I've reserved KIP number 444 for it) would contain >> all the changes on StreamsPartitionAssginor to allow warming up new >> members. >> >> I think the first part, a.k.a. the current updated KIP-429 is ready for >> review and discussions again. Would love to hear people's feedbacks and >> ideas. >> >> Guozhang >> >> >> >> On Mon, Mar 4, 2019 at 10:09 AM Boyang Chen wrote: >> >>> Thanks Guozhang for the great questions. Answers are inlined: >>> >>> 1. I'm still not sure if it's worthwhile to add a new type of "learner >>> task" in addition to "standby task": if the only difference is that for >>> the >>> latter, we would consider workload balance while for the former we would >>> not, I think we can just adjust the logic of StickyTaskAssignor a bit to >>> break that difference. Adding a new type of task would be adding a lot of >>> code complexity, so if we can still piggy-back the logic on a standby-task >>> I would prefer to do so. >>> In the proposal we stated that we are not adding a new type of task >>> implementation. The >>> learner task shall share the same implementation with normal standby >>> task, only that we >>> shall tag the st
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
ose default implementation is >>> StickyPartitionAssignor. Streams partition assignor logic today sites in >>> the latter two classes. Hence the hierarchy today is: >>> >>> KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> >>> StickyTaskAssignor. >>> >>> We need to think about where the proposed implementation would take place >>> at, and personally I think it is not the best option to inject all of them >>> into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of >>> "triggering another rebalance" etc would require some coordinator logic >>> which is hard to mimic at PartitionAssignor level. On the other hand, >>> since >>> we are embedding a KafkaConsumer client as a whole we cannot just replace >>> ConsumerCoordinator with a specialized StreamsCoordinator like Connect >>> does >>> in KIP-415. So I'd like to maybe split the current proposal in both >>> consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. >>> And then the key thing to consider is how to cut off the boundary so that >>> the modifications we push to ConsumerCoordinator would be beneficial >>> universally for any consumers, while keep the Streams-specific logic at >>> the >>> assignor level. >>> Yes, that's also my ideal plan. The details for the implementation are >>> depicted >>> in this doc< >>> https://docs.google.com/document/d/1me2a5wvxAZT1QE6HkwyDl7C2TiBQlKN3Dpw_I1ro91U/edit#heading=h.qix74qdmekae>, >>> and I have explained the reasoning on why we want to push a >>> global change of replacing ConsumerCoordinator with StreamCoordinator. >>> The motivation >>> is that KIP space is usually used for public & algorithm level change, >>> not for internal >>> implementation details. >>> >>> 3. Depending on which design direction we choose, our migration plan would >>> also be quite different. For example, if we stay with ConsumerCoordinator >>> whose protocol type is "consumer" still, and we can manage to make all >>> changes agnostic to brokers as well as to old versioned consumers, then >>> our >>> migration plan could be much easier. >>> Yes, the upgrade plan was designed to take the new StreamCoordinator >>> approach >>> which means we shall define a new protocol type. For existing application >>> we could only >>> maintain the same `consumer` protocol type is because current broker only >>> allows >>> change of protocol type when the consumer group is empty. It is of course >>> user-unfriendly to force >>> a wipe-out for the entire application, and I don't think maintaining old >>> protocol type would greatly >>> impact ongoing services using new stream coordinator. WDYT? >>> >>> 4. I think one major issue related to this KIP is that today, in the >>> StickyPartitionAssignor, we always try to honor stickiness over workload >>> balance, and hence "learner task" is needed to break this priority, but >>> I'm >>> wondering if we can have a better solution within sticky task assignor >>> that >>> accommodate this? >>> Great question! That's what I explained in the proposal, which is that we >>> should breakdown our >>> delivery into different stages. At very beginning, our goal is to trigger >>> learner task assignment only on >>> `new` hosts, where we shall leverage leader's knowledge of previous round >>> of rebalance to figure out. After >>> stage one, our goal is to have a smooth scaling up experience, but the >>> task balance problem is kind of orthogonal. >>> The load balance problem is a much broader topic than auto scaling, which >>> I figure worth discussing within >>> this KIP's context since it's a naturally next-step, but wouldn't be the >>> main topic. >>> Learner task or auto scaling support should be treated as `a helpful >>> mechanism to reach load balance`, but not `an algorithm defining load >>> balance`. It would be great if you could share some insights of the stream >>> task balance, which eventually helps us to break out of the KIP-429's scope >>> and even define a separate KIP to focus on task weight & assignment logic >>> improvement. >>> >>> Also thank you for making improvement on the KIP context and organization! >>> >>> Best, >>> Boyang >>> >>>
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
pproach >> which means we shall define a new protocol type. For existing application >> we could only >> maintain the same `consumer` protocol type is because current broker only >> allows >> change of protocol type when the consumer group is empty. It is of course >> user-unfriendly to force >> a wipe-out for the entire application, and I don't think maintaining old >> protocol type would greatly >> impact ongoing services using new stream coordinator. WDYT? >> >> 4. I think one major issue related to this KIP is that today, in the >> StickyPartitionAssignor, we always try to honor stickiness over workload >> balance, and hence "learner task" is needed to break this priority, but >> I'm >> wondering if we can have a better solution within sticky task assignor >> that >> accommodate this? >> Great question! That's what I explained in the proposal, which is that we >> should breakdown our >> delivery into different stages. At very beginning, our goal is to trigger >> learner task assignment only on >> `new` hosts, where we shall leverage leader's knowledge of previous round >> of rebalance to figure out. After >> stage one, our goal is to have a smooth scaling up experience, but the >> task balance problem is kind of orthogonal. >> The load balance problem is a much broader topic than auto scaling, which >> I figure worth discussing within >> this KIP's context since it's a naturally next-step, but wouldn't be the >> main topic. >> Learner task or auto scaling support should be treated as `a helpful >> mechanism to reach load balance`, but not `an algorithm defining load >> balance`. It would be great if you could share some insights of the stream >> task balance, which eventually helps us to break out of the KIP-429's scope >> and even define a separate KIP to focus on task weight & assignment logic >> improvement. >> >> Also thank you for making improvement on the KIP context and organization! >> >> Best, >> Boyang >> >> From: Guozhang Wang >> Sent: Saturday, March 2, 2019 6:00 AM >> To: dev >> Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams >> >> Hello Boyang, >> >> I've just made a quick pass on the KIP and here are some thoughts. >> >> Meta: >> >> 1. I'm still not sure if it's worthwhile to add a new type of "learner >> task" in addition to "standby task": if the only difference is that for >> the >> latter, we would consider workload balance while for the former we would >> not, I think we can just adjust the logic of StickyTaskAssignor a bit to >> break that difference. Adding a new type of task would be adding a lot of >> code complexity, so if we can still piggy-back the logic on a standby-task >> I would prefer to do so. >> >> 2. One thing that's still not clear from the KIP wiki itself is which >> layer >> would the logic be implemented at. Although for most KIPs we would not >> require internal implementation details but only public facing API >> updates, >> for a KIP like this I think it still requires to flesh out details on the >> implementation design. More specifically: today Streams embed a full >> fledged Consumer client, which hard-code a ConsumerCoordinator inside, >> Streams then injects a StreamsPartitionAssignor to its plugable >> PartitionAssignor interface and inside the StreamsPartitionAssignor we >> also >> have a TaskAssignor interface whose default implementation is >> StickyPartitionAssignor. Streams partition assignor logic today sites in >> the latter two classes. Hence the hierarchy today is: >> >> KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> >> StickyTaskAssignor. >> >> We need to think about where the proposed implementation would take place >> at, and personally I think it is not the best option to inject all of them >> into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of >> "triggering another rebalance" etc would require some coordinator logic >> which is hard to mimic at PartitionAssignor level. On the other hand, >> since >> we are embedding a KafkaConsumer client as a whole we cannot just replace >> ConsumerCoordinator with a specialized StreamsCoordinator like Connect >> does >> in KIP-415. So I'd like to maybe split the current proposal in both >> consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. >> And then the key thing to consider is how to cut off the boundar
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Sounds good Guozhang! I will kick another discussion thread for KIP-444. Boyang From: Guozhang Wang Sent: Thursday, March 21, 2019 9:17 AM To: dev Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams Hello Boyang, I've made another thorough pass over this KIP and I'd like to spilt it into two parts: the first part, covered in KIP-429 would be touching on Consumer Coordinator only to have incremental rebalance protocol in place. The second part (for now I've reserved KIP number 444 for it) would contain all the changes on StreamsPartitionAssginor to allow warming up new members. I think the first part, a.k.a. the current updated KIP-429 is ready for review and discussions again. Would love to hear people's feedbacks and ideas. Guozhang On Mon, Mar 4, 2019 at 10:09 AM Boyang Chen wrote: > Thanks Guozhang for the great questions. Answers are inlined: > > 1. I'm still not sure if it's worthwhile to add a new type of "learner > task" in addition to "standby task": if the only difference is that for the > latter, we would consider workload balance while for the former we would > not, I think we can just adjust the logic of StickyTaskAssignor a bit to > break that difference. Adding a new type of task would be adding a lot of > code complexity, so if we can still piggy-back the logic on a standby-task > I would prefer to do so. > In the proposal we stated that we are not adding a new type of task > implementation. The > learner task shall share the same implementation with normal standby task, > only that we > shall tag the standby task with learner and prioritize the learner tasks > replay effort. > 2. One thing that's still not clear from the KIP wiki itself is which layer > would the logic be implemented at. Although for most KIPs we would not > require internal implementation details but only public facing API updates, > for a KIP like this I think it still requires to flesh out details on the > implementation design. More specifically: today Streams embed a full > fledged Consumer client, which hard-code a ConsumerCoordinator inside, > Streams then injects a StreamsPartitionAssignor to its pluggable > PartitionAssignor interface and inside the StreamsPartitionAssignor we also > have a TaskAssignor interface whose default implementation is > StickyPartitionAssignor. Streams partition assignor logic today sites in > the latter two classes. Hence the hierarchy today is: > > KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> > StickyTaskAssignor. > > We need to think about where the proposed implementation would take place > at, and personally I think it is not the best option to inject all of them > into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of > "triggering another rebalance" etc would require some coordinator logic > which is hard to mimic at PartitionAssignor level. On the other hand, since > we are embedding a KafkaConsumer client as a whole we cannot just replace > ConsumerCoordinator with a specialized StreamsCoordinator like Connect does > in KIP-415. So I'd like to maybe split the current proposal in both > consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. > And then the key thing to consider is how to cut off the boundary so that > the modifications we push to ConsumerCoordinator would be beneficial > universally for any consumers, while keep the Streams-specific logic at the > assignor level. > Yes, that's also my ideal plan. The details for the implementation are > depicted > in this doc< > https://docs.google.com/document/d/1me2a5wvxAZT1QE6HkwyDl7C2TiBQlKN3Dpw_I1ro91U/edit#heading=h.qix74qdmekae>, > and I have explained the reasoning on why we want to push a > global change of replacing ConsumerCoordinator with StreamCoordinator. The > motivation > is that KIP space is usually used for public & algorithm level change, not > for internal > implementation details. > > 3. Depending on which design direction we choose, our migration plan would > also be quite different. For example, if we stay with ConsumerCoordinator > whose protocol type is "consumer" still, and we can manage to make all > changes agnostic to brokers as well as to old versioned consumers, then our > migration plan could be much easier. > Yes, the upgrade plan was designed to take the new StreamCoordinator > approach > which means we shall define a new protocol type. For existing application > we could only > maintain the same `consumer` protocol type is because current broker only > allows > change of protocol type when the consumer group is empty. It is of course > user-unfriendly to force > a wipe-out for the entire application, and I don'
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
ess over workload > balance, and hence "learner task" is needed to break this priority, but I'm > wondering if we can have a better solution within sticky task assignor that > accommodate this? > Great question! That's what I explained in the proposal, which is that we > should breakdown our > delivery into different stages. At very beginning, our goal is to trigger > learner task assignment only on > `new` hosts, where we shall leverage leader's knowledge of previous round > of rebalance to figure out. After > stage one, our goal is to have a smooth scaling up experience, but the > task balance problem is kind of orthogonal. > The load balance problem is a much broader topic than auto scaling, which > I figure worth discussing within > this KIP's context since it's a naturally next-step, but wouldn't be the > main topic. > Learner task or auto scaling support should be treated as `a helpful > mechanism to reach load balance`, but not `an algorithm defining load > balance`. It would be great if you could share some insights of the stream > task balance, which eventually helps us to break out of the KIP-429's scope > and even define a separate KIP to focus on task weight & assignment logic > improvement. > > Also thank you for making improvement on the KIP context and organization! > > Best, > Boyang > > From: Guozhang Wang > Sent: Saturday, March 2, 2019 6:00 AM > To: dev > Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams > > Hello Boyang, > > I've just made a quick pass on the KIP and here are some thoughts. > > Meta: > > 1. I'm still not sure if it's worthwhile to add a new type of "learner > task" in addition to "standby task": if the only difference is that for the > latter, we would consider workload balance while for the former we would > not, I think we can just adjust the logic of StickyTaskAssignor a bit to > break that difference. Adding a new type of task would be adding a lot of > code complexity, so if we can still piggy-back the logic on a standby-task > I would prefer to do so. > > 2. One thing that's still not clear from the KIP wiki itself is which layer > would the logic be implemented at. Although for most KIPs we would not > require internal implementation details but only public facing API updates, > for a KIP like this I think it still requires to flesh out details on the > implementation design. More specifically: today Streams embed a full > fledged Consumer client, which hard-code a ConsumerCoordinator inside, > Streams then injects a StreamsPartitionAssignor to its plugable > PartitionAssignor interface and inside the StreamsPartitionAssignor we also > have a TaskAssignor interface whose default implementation is > StickyPartitionAssignor. Streams partition assignor logic today sites in > the latter two classes. Hence the hierarchy today is: > > KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> > StickyTaskAssignor. > > We need to think about where the proposed implementation would take place > at, and personally I think it is not the best option to inject all of them > into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of > "triggering another rebalance" etc would require some coordinator logic > which is hard to mimic at PartitionAssignor level. On the other hand, since > we are embedding a KafkaConsumer client as a whole we cannot just replace > ConsumerCoordinator with a specialized StreamsCoordinator like Connect does > in KIP-415. So I'd like to maybe split the current proposal in both > consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. > And then the key thing to consider is how to cut off the boundary so that > the modifications we push to ConsumerCoordinator would be beneficial > universally for any consumers, while keep the Streams-specific logic at the > assignor level. > > 3. Depending on which design direction we choose, our migration plan would > also be quite different. For example, if we stay with ConsumerCoordinator > whose protocol type is "consumer" still, and we can manage to make all > changes agnostic to brokers as well as to old versioned consumers, then our > migration plan could be much easier. > > 4. I think one major issue related to this KIP is that today, in the > StickyPartitionAssignor, we always try to honor stickiness over workload > balance, and hence "learner task" is needed to break this priority, but I'm > wondering if we can have a better solution within sticky task assignor that > accommodate this? > > Minor: > > 1. The idea of two rebalances have also been discussed in > https://iss
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
ng load balance`. It would be great if you could share some insights of the stream task balance, which eventually helps us to break out of the KIP-429's scope and even define a separate KIP to focus on task weight & assignment logic improvement. Also thank you for making improvement on the KIP context and organization! Best, Boyang From: Guozhang Wang Sent: Saturday, March 2, 2019 6:00 AM To: dev Subject: Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams Hello Boyang, I've just made a quick pass on the KIP and here are some thoughts. Meta: 1. I'm still not sure if it's worthwhile to add a new type of "learner task" in addition to "standby task": if the only difference is that for the latter, we would consider workload balance while for the former we would not, I think we can just adjust the logic of StickyTaskAssignor a bit to break that difference. Adding a new type of task would be adding a lot of code complexity, so if we can still piggy-back the logic on a standby-task I would prefer to do so. 2. One thing that's still not clear from the KIP wiki itself is which layer would the logic be implemented at. Although for most KIPs we would not require internal implementation details but only public facing API updates, for a KIP like this I think it still requires to flesh out details on the implementation design. More specifically: today Streams embed a full fledged Consumer client, which hard-code a ConsumerCoordinator inside, Streams then injects a StreamsPartitionAssignor to its plugable PartitionAssignor interface and inside the StreamsPartitionAssignor we also have a TaskAssignor interface whose default implementation is StickyPartitionAssignor. Streams partition assignor logic today sites in the latter two classes. Hence the hierarchy today is: KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> StickyTaskAssignor. We need to think about where the proposed implementation would take place at, and personally I think it is not the best option to inject all of them into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of "triggering another rebalance" etc would require some coordinator logic which is hard to mimic at PartitionAssignor level. On the other hand, since we are embedding a KafkaConsumer client as a whole we cannot just replace ConsumerCoordinator with a specialized StreamsCoordinator like Connect does in KIP-415. So I'd like to maybe split the current proposal in both consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. And then the key thing to consider is how to cut off the boundary so that the modifications we push to ConsumerCoordinator would be beneficial universally for any consumers, while keep the Streams-specific logic at the assignor level. 3. Depending on which design direction we choose, our migration plan would also be quite different. For example, if we stay with ConsumerCoordinator whose protocol type is "consumer" still, and we can manage to make all changes agnostic to brokers as well as to old versioned consumers, then our migration plan could be much easier. 4. I think one major issue related to this KIP is that today, in the StickyPartitionAssignor, we always try to honor stickiness over workload balance, and hence "learner task" is needed to break this priority, but I'm wondering if we can have a better solution within sticky task assignor that accommodate this? Minor: 1. The idea of two rebalances have also been discussed in https://issues.apache.org/jira/browse/KAFKA-6145. So we should add the reference on the wiki page as well. 2. Could you also add a section describing how the subscription / assignment metadata will be re-formatted? Without this information it is hard to get to the bottom of your idea. For example in the "Leader Transfer Before Scaling" section, I'm not sure why "S2 doesn't know S4 is new member" and hence would blindly obey stickiness over workload balance requirement. Guozhang On Thu, Feb 28, 2019 at 11:05 AM Boyang Chen wrote: > Hey community friends, > > I'm gladly inviting you to have a look at the proposal to add incremental > rebalancing to Kafka Streams, A.K.A auto-scaling support. > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Smooth+Auto-Scaling+for+Kafka+Streams > > Special thanks to Guozhang for giving great guidances and important > feedbacks while making this KIP! > > Best, > Boyang > -- -- Guozhang
Re: [DISCUSS] KIP-429 : Smooth Auto-Scaling for Kafka Streams
Hello Boyang, I've just made a quick pass on the KIP and here are some thoughts. Meta: 1. I'm still not sure if it's worthwhile to add a new type of "learner task" in addition to "standby task": if the only difference is that for the latter, we would consider workload balance while for the former we would not, I think we can just adjust the logic of StickyTaskAssignor a bit to break that difference. Adding a new type of task would be adding a lot of code complexity, so if we can still piggy-back the logic on a standby-task I would prefer to do so. 2. One thing that's still not clear from the KIP wiki itself is which layer would the logic be implemented at. Although for most KIPs we would not require internal implementation details but only public facing API updates, for a KIP like this I think it still requires to flesh out details on the implementation design. More specifically: today Streams embed a full fledged Consumer client, which hard-code a ConsumerCoordinator inside, Streams then injects a StreamsPartitionAssignor to its plugable PartitionAssignor interface and inside the StreamsPartitionAssignor we also have a TaskAssignor interface whose default implementation is StickyPartitionAssignor. Streams partition assignor logic today sites in the latter two classes. Hence the hierarchy today is: KafkaConsumer -> ConsumerCoordinator -> StreamsPartitionAssignor -> StickyTaskAssignor. We need to think about where the proposed implementation would take place at, and personally I think it is not the best option to inject all of them into the StreamsPartitionAssignor / StickyTaskAssignor since the logic of "triggering another rebalance" etc would require some coordinator logic which is hard to mimic at PartitionAssignor level. On the other hand, since we are embedding a KafkaConsumer client as a whole we cannot just replace ConsumerCoordinator with a specialized StreamsCoordinator like Connect does in KIP-415. So I'd like to maybe split the current proposal in both consumer layer and streams-assignor layer like we did in KIP-98/KIP-129. And then the key thing to consider is how to cut off the boundary so that the modifications we push to ConsumerCoordinator would be beneficial universally for any consumers, while keep the Streams-specific logic at the assignor level. 3. Depending on which design direction we choose, our migration plan would also be quite different. For example, if we stay with ConsumerCoordinator whose protocol type is "consumer" still, and we can manage to make all changes agnostic to brokers as well as to old versioned consumers, then our migration plan could be much easier. 4. I think one major issue related to this KIP is that today, in the StickyPartitionAssignor, we always try to honor stickiness over workload balance, and hence "learner task" is needed to break this priority, but I'm wondering if we can have a better solution within sticky task assignor that accommodate this? Minor: 1. The idea of two rebalances have also been discussed in https://issues.apache.org/jira/browse/KAFKA-6145. So we should add the reference on the wiki page as well. 2. Could you also add a section describing how the subscription / assignment metadata will be re-formatted? Without this information it is hard to get to the bottom of your idea. For example in the "Leader Transfer Before Scaling" section, I'm not sure why "S2 doesn't know S4 is new member" and hence would blindly obey stickiness over workload balance requirement. Guozhang On Thu, Feb 28, 2019 at 11:05 AM Boyang Chen wrote: > Hey community friends, > > I'm gladly inviting you to have a look at the proposal to add incremental > rebalancing to Kafka Streams, A.K.A auto-scaling support. > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Smooth+Auto-Scaling+for+Kafka+Streams > > Special thanks to Guozhang for giving great guidances and important > feedbacks while making this KIP! > > Best, > Boyang > -- -- Guozhang
