THIS IS A TEST INSTANCE. ALL YOUR CHANGES WILL BE LOST!!!!
...
We will bump the OffsetFetch API. The new schemas are provided below. These introduce a group_id level return, note that previous top level error codes will now be return at a per-group level:
| Code Block |
|---|
OffsetFetch Request => [group_ids] require_stable TAG_BUFFER
group_ids => name [topics] TAG_BUFFER
topics => name [partition_indexes] TAG_BUFFER
name => COMPACT_STRING
partition_indexes => INT32
require_stable => BOOLEAN
OffsetFetch Response => throttle_time_ms [group_ids] TAG_BUFFER
throttle_time_ms => INT32
group_ids => name [topics] TAG_BUFFER // new
topics => name [partitions] TAG_BUFFER
name => COMPACT_STRING
partitions => partition_index committed_offset committed_leader_epoch metadata error_code TAG_BUFFER
partition_index => INT32
committed_offset => INT64
committed_leader_epoch => INT32
metadata => COMPACT_NULLABLE_STRING
error_code => INT16
error_code => INT16 // moved |
The new level of groups information in the response is reflected in OffsetFetchResponseData
...
The changes contained in this KIP refer specifically to offset listing requests. However, in order to perform these requests clients must first find the correct group coordinator. Optimisations for this process are contained in KIP-699 (KIP-699: FindCoordinatorsUpdate FindCoordinator to resolve multiple Coordinators at a time). These 2 KIPS are not interdependent.
...