[ 
https://issues.apache.org/jira/browse/KAFKA-20623?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Alieh Saeedi updated KAFKA-20623:
---------------------------------
    Description: 
Write path, heartbeat gating, delete:



GroupCoordinatorService: builder plumbing for plugin. Heartbeat 
post-processing: set TopologyDescriptionRequired=true when StoredTopologyEpoch 
!= currentTopologyEpoch AND LastFailedTopologyEpoch != currentTopologyEpoch AND 
backoff not active (skip STALE_TOPOLOGY members). In-memory per-group 
exponential backoff (30s → 1h). StreamsGroupTopologyDescriptionUpdate handler: 
validate MemberId against current group membership (→ UNKNOWN_MEMBER_ID), call 
plugin.setTopology, on success write StoredTopologyEpoch metadata record, on 
permanent failure write LastFailedTopologyEpoch, on transient failure arm 
backoff. POJO↔wire converters. deleteGroups integration: call 
plugin.deleteTopology before tombstoning, return DELETE_FAILED on plugin 
failure. GroupCoordinator interface update. GroupCoordinatorServiceTest.

> Wire topology description plugin into GroupCoordinatorService
> -------------------------------------------------------------
>
>                 Key: KAFKA-20623
>                 URL: https://issues.apache.org/jira/browse/KAFKA-20623
>             Project: Kafka
>          Issue Type: Sub-task
>            Reporter: Alieh Saeedi
>            Assignee: Lucas Brutschy
>            Priority: Major
>             Fix For: 4.4.0
>
>
> Write path, heartbeat gating, delete:
> GroupCoordinatorService: builder plumbing for plugin. Heartbeat 
> post-processing: set TopologyDescriptionRequired=true when 
> StoredTopologyEpoch != currentTopologyEpoch AND LastFailedTopologyEpoch != 
> currentTopologyEpoch AND backoff not active (skip STALE_TOPOLOGY members). 
> In-memory per-group exponential backoff (30s → 1h). 
> StreamsGroupTopologyDescriptionUpdate handler: validate MemberId against 
> current group membership (→ UNKNOWN_MEMBER_ID), call plugin.setTopology, on 
> success write StoredTopologyEpoch metadata record, on permanent failure write 
> LastFailedTopologyEpoch, on transient failure arm backoff. POJO↔wire 
> converters. deleteGroups integration: call plugin.deleteTopology before 
> tombstoning, return DELETE_FAILED on plugin failure. GroupCoordinator 
> interface update. GroupCoordinatorServiceTest.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to