[
https://issues.apache.org/jira/browse/KAFKA-12463?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17308235#comment-17308235
]
Chris Egerton commented on KAFKA-12463:
---------------------------------------
Thanks [~kkonstantine]--I'll admit I was hoping to avoid a KIP as this seems
like more of an implementation detail and not something most users would
consider when using Connect, but the longer the conversation has gone on it's
become clear that this is a measure-twice, cut-once situation with multiple
perspectives and tradeoffs that need to be taken into account. Regardless of
whether a KIP is strictly required or not, I certainly agree it'd be best at
this point to proceed with one, so I'll add the {{needs-kip}} tag and will
hopefully get a design doc drafted and on the mailing list within the next week
or two.
> Update default consumer partition assignor for sink tasks
> ---------------------------------------------------------
>
> Key: KAFKA-12463
> URL: https://issues.apache.org/jira/browse/KAFKA-12463
> Project: Kafka
> Issue Type: Improvement
> Components: KafkaConnect
> Reporter: Chris Egerton
> Assignee: Chris Egerton
> Priority: Major
> Labels: needs-kip
>
> Kafka consumers have a pluggable [partition assignment
> interface|https://kafka.apache.org/27/javadoc/org/apache/kafka/clients/consumer/ConsumerPartitionAssignor.html]
> that comes with several out-of-the-box implementations including the
> [RangeAssignor|https://kafka.apache.org/27/javadoc/org/apache/kafka/clients/consumer/RangeAssignor.html],
>
> [RoundRobinAssignor|https://kafka.apache.org/27/javadoc/org/apache/kafka/clients/consumer/RoundRobinAssignor.html],
>
> [StickyAssignor|https://kafka.apache.org/27/javadoc/org/apache/kafka/clients/consumer/StickyAssignor.html],
> and
> [CooperativeStickyAssignor|https://kafka.apache.org/27/javadoc/org/apache/kafka/clients/consumer/CooperativeStickyAssignor.html].
> If no partition assignor is configured with a consumer, the {{RangeAssignor}}
> is used by default. Although there are some benefits to this assignor
> including stability of assignment across generations and simplicity of
> design, it comes with a major drawback: the number of active consumers in a
> group is limited to the number of partitions in the topic(s) with the most
> partitions. For an example of the worst case, in a consumer group where every
> member is subscribed to ten topics that each have one partition, only one
> member of that group will be assigned any topic partitions.
> This can end up producing counterintuitive and even frustrating behavior when
> a sink connector is brought up with N tasks to read from some collection of
> topics with a total of N topic partitions, but some tasks end up idling and
> not processing any data.
> h3. Proposed Change
> *NOTE: Until/unless KAFKA-12477 is addressed, the approach outlined below
> will not work as consumers will still perform eager rebalancing as long as at
> least one of the partition assignors they are configured with does not
> support cooperative rebalancing. KAFKA-12487 should also be addressed before
> configuring any Connect worker to use the {{CooperativeStickyAssignor}} for
> any sink connectors.*
> [KIP-429|https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol]
> introduced the {{CooperativeStickyAssignor}}, which seeks to provide a
> stable assignment across generations wherever possible, provide the most even
> assignment possible (taking into account possible differences in
> subscriptions across consumers in the group), and allow consumers to continue
> processing data during rebalance. The documentation for the assignor states
> that "Users should prefer this assignor for newer clusters." As Connect and
> the tooling around it matures and automatic restarts of failed tasks become
> more popular, care should be taken to ensure that the consumer group churn
> created by restarting one or more tasks doesn't compromise the availability
> of other tasks by forcing them to temporarily yield up all of their
> partitions just to reclaim them after a rebalance has completed.
> With that in mind, we should alter the default consumer configuration for
> sink tasks to use the new {{CooperativeStickyAssignor}}. In order to do this
> in a backwards-compatible fashion that also enables rolling upgrades, this
> should be implemented by changing the {{Worker}} to set the following on the
> consumer configuration created for each sink connector task:
> {code:java}
> partition.assignment.strategy=org.apache.kafka.clients.consumer.CooperativeStickyAssignor,org.apache.kafka.clients.consumer.RangeAssignor
> {code}
> This way, consumer groups for sink connectors on Connect clusters in the
> process of being upgraded will continue to use the {{RangeAssignor}} until
> all workers in the cluster have been upgraded, and then will switch over to
> the new {{CooperativeStickyAssignor}} automatically.
> Importantly, this setting will only be a default, and any user-specified
> overrides either in the *worker config*:
>
> {code:java}
> consumer.partition.assignment.strategy=<user-specified strategy>{code}
>
> or in the *connector config*:
>
> {code:java}
> "consumer.override.partition.assignment.strategy": "<user-specified
> strategy>"{code}
>
> will still be respected.
> This improvement is viable as far back as -2.3- 2.4, when the
> {{CooperativeStickyAssignor}} was introduced, but given that it is not a bug
> fix, should only be applied to the Connect framework in an upcoming minor
> release.
> h3. Workaround: manually setting the partition assignment strategy
> There is a simple workaround to achieve the same behavior in releases 2.4 and
> later that don't include this improvement: manually override either a
> connector configuration or an entire worker configuration.
> In order to avoid task failures while the connector is being reconfigured, it
> is highly recommended that the consumer be configured with a list of both the
> new and the current partition assignment strategies, instead of just the new
> partition assignment strategy.
>
> For example, to update a connector formerly configured to use the
> {{RangeAssignor}} strategy to instead use the {{RoundRobinAssignor}}
> strategy, add the following to the connector configuration:
> {code:java}
> "consumer.override.partition.assignment.strategy":
> "org.apache.kafka.clients.consumer.RoundRobinAssignor,
> org.apache.kafka.clients.consumer.RangeAssignor"{code}
> Note that this will require per-connector client overrides to be enabled on
> the worker, which can be accomplished by including this in the worker's
> configuration:
> {code:java}
> connector.client.config.override.policy=ALL
> {code}
> And to update an entire worker formerly configured to use the
> {{RangeAssignor}} strategy to instead use the {{RoundRobinAssignor}}
> strategy, add the following to the worker configuration:
> {code:java}
> consumer.partition.assignment.strategy=org.apache.kafka.clients.consumer.RoundRobinAssignor,
> org.apache.kafka.clients.consumer.RangeAssignor{code}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
