[
https://issues.apache.org/jira/browse/KAFKA-12495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17582710#comment-17582710
]
Sagar Rao commented on KAFKA-12495:
-----------------------------------
Thanks [~showuon] for your comments.
Regarding the new draft KIP thank you for giving a pass through it. I am
working on a new version based on initial suggestions/feedback. As you pointed
out, it keeps the assignment logic at the client side - similar to streams and
the current rebalancing protocol. Yeah it could very well be moved to the group
coordinator layer.
Having said that, I decided to keep it in the client layer because in
Incremental Cooperative protocol, clients can trigger a rebalance based on the
scheduled rebalance delay. As per my understanding of KIP-848, such rebalances
which the clients want to trigger, can be done by setting the `reason` field.
That kind of rebalance triggering is similar(yet not as complex) as the kind of
rebalance that Streams triggers. Let me know if that makes sense.
Also, it would be great if this feedback can be shared on the discussion
thread! It would evoke further comments from more members in the community
which would help in solidifying the design! Thanks!
> Unbalanced connectors/tasks distribution will happen in Connect's incremental
> cooperative assignor
> --------------------------------------------------------------------------------------------------
>
> Key: KAFKA-12495
> URL: https://issues.apache.org/jira/browse/KAFKA-12495
> Project: Kafka
> Issue Type: Bug
> Components: KafkaConnect
> Reporter: Luke Chen
> Assignee: Sagar Rao
> Priority: Critical
> Attachments: image-2021-03-18-15-04-57-854.png,
> image-2021-03-18-15-05-52-557.png, image-2021-03-18-15-07-27-103.png
>
>
> In Kafka Connect, we implement incremental cooperative rebalance algorithm
> based on KIP-415
> ([https://cwiki.apache.org/confluence/display/KAFKA/KIP-415%3A+Incremental+Cooperative+Rebalancing+in+Kafka+Connect)|https://cwiki.apache.org/confluence/display/KAFKA/KIP-415%3A+Incremental+Cooperative+Rebalancing+in+Kafka+Connect].
> However, we have a bad assumption in the algorithm implementation, which is:
> after revoking rebalance completed, the member(worker) count will be the same
> as the previous round of reblance.
>
> Let's take a look at the example in the KIP-415:
> !image-2021-03-18-15-07-27-103.png|width=441,height=556!
> It works well for most cases. But what if W4 added after 1st rebalance
> completed and before 2nd rebalance started? Let's see what will happened?
> Let's see this example: (we'll use 10 tasks here):
>
> {code:java}
> Initial group and assignment: W1([AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1,
> BT2, BT4, BT4, BT5])
> Config topic contains: AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1, BT2, BT4, BT4,
> BT5
> W1 is current leader
> W2 joins with assignment: []
> Rebalance is triggered
> W3 joins while rebalance is still active with assignment: []
> W1 joins with assignment: [AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1, BT2, BT4,
> BT4, BT5]
> W1 becomes leader
> W1 computes and sends assignments:
> W1(delay: 0, assigned: [AC0, AT1, AT2, AT3], revoked: [AT4, AT5, BC0, BT1,
> BT2, BT4, BT4, BT5])
> W2(delay: 0, assigned: [], revoked: [])
> W3(delay: 0, assigned: [], revoked: [])
> W1 stops revoked resources
> W1 rejoins with assignment: [AC0, AT1, AT2, AT3]
> Rebalance is triggered
> W2 joins with assignment: []
> W3 joins with assignment: []
> // one more member joined
> W4 joins with assignment: []
> W1 becomes leader
> W1 computes and sends assignments:
> // We assigned all the previous revoked Connectors/Tasks to the new member,
> but we didn't revoke any more C/T in this round, which cause unbalanced
> distribution
> W1(delay: 0, assigned: [AC0, AT1, AT2, AT3], revoked: [])
> W2(delay: 0, assigned: [AT4, AT5, BC0], revoked: [])
> W2(delay: 0, assigned: [BT1, BT2, BT4], revoked: [])
> W2(delay: 0, assigned: [BT4, BT5], revoked: [])
> {code}
> Because we didn't allow to do consecutive revoke in two consecutive
> rebalances (under the same leader), we will have this uneven distribution
> under this situation. We should allow consecutive rebalance to have another
> round of revocation to revoke the C/T to the other members in this case.
> expected:
> {code:java}
> Initial group and assignment: W1([AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1,
> BT2, BT4, BT4, BT5])
> Config topic contains: AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1, BT2, BT4, BT4,
> BT5
> W1 is current leader
> W2 joins with assignment: []
> Rebalance is triggered
> W3 joins while rebalance is still active with assignment: []
> W1 joins with assignment: [AC0, AT1, AT2, AT3, AT4, AT5, BC0, BT1, BT2, BT4,
> BT4, BT5]
> W1 becomes leader
> W1 computes and sends assignments:
> W1(delay: 0, assigned: [AC0, AT1, AT2, AT3], revoked: [AT4, AT5, BC0, BT1,
> BT2, BT4, BT4, BT5])
> W2(delay: 0, assigned: [], revoked: [])
> W3(delay: 0, assigned: [], revoked: [])
> W1 stops revoked resources
> W1 rejoins with assignment: [AC0, AT1, AT2, AT3]
> Rebalance is triggered
> W2 joins with assignment: []
> W3 joins with assignment: []
> // one more member joined
> W4 joins with assignment: []
> W1 becomes leader
> W1 computes and sends assignments:
> // We assigned all the previous revoked Connectors/Tasks to the new member,
> **and also revoke some C/T**
> W1(delay: 0, assigned: [AC0, AT1, AT2], revoked: [AT3])
> W2(delay: 0, assigned: [AT4, AT5, BC0], revoked: [])
> W3(delay: 0, assigned: [BT1, BT2, BT4], revoked: [])
> W4(delay: 0, assigned: [BT4, BT5], revoked: [])
> // another round of rebalance to assign the new revoked C/T to the other
> members
> W1 rejoins with assignment: [AC0, AT1, AT2]
> Rebalance is triggered
> W2 joins with assignment: [AT4, AT5, BC0]
> W3 joins with assignment: [BT1, BT2, BT4]
> W4 joins with assignment: [BT4, BT5]
> W1 becomes leader
> W1 computes and sends assignments:
> // (final) We assigned all the previous revoked Connectors/Tasks to the
> members
> W1(delay: 0, assigned: [AC0, AT1, AT2], revoked: [])
> W2(delay: 0, assigned: [AT4, AT5, BC0], revoked: [])
> W2(delay: 0, assigned: [BT1, BT2, BT4], revoked: [])
> W2(delay: 0, assigned: [BT4, BT5, AT3], revoked: [])
> {code}
> Note: The consumer's cooperative sticky assignor won't have this issue since
> we re-compute the assignment in each round.
>
> Note2: this issue makes KAFKA-12283 test flaky.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
