[
https://issues.apache.org/jira/browse/KAFKA-16032?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lianet Magrans updated KAFKA-16032:
-----------------------------------
Description:
OffsetFetch and OffsetCommit handle errors separately. There are 2 issues to
review around this:
- The logic is duplicated for some errors that are treated similarly (ex.
NOT_COORDINATOR)
- Some errors are not handled similarly in both requests (ex.
COORDINATOR_NOT_AVAILABLE handled and retried for OffsetCommit but not
OffsetFetch). Note that the specific errors handled by each request were kept
the same as in the legacy ConsumerCoordinator but this should be reviewed, in
an attempt to handle the same errors, in the same way, whenever possible.
Note that the legacy approach handles expected errors for each path (FETCH and
COMMIT), retrying on those when needed, but does not retry on unexpected
retriable errors.
was:
OffsetFetch and OffsetCommit handle errors separately. There are 2 issues to
review around this:
- The logic is duplicated for some errors that are treated similarly (ex.
NOT_COORDINATOR)
- Some errors are not handled similarly in both requests (ex.
COORDINATOR_NOT_AVAILABLE handled and retried for OffsetCommit but not
OffsetFetch). Note that the specific errors handled by each request were kept
the same as in the legacy ConsumerCoordinator but this should be reviewed, in
an attempt to handle the same errors, in the same way, whenever possible.
This should be reviewed also with the goal of unifying the retry logic around
those errors that is now applies in multiple different places of the manager
depending on the request path.
> Review inconsistent error handling of OffsetFetch and OffsetCommit responses
> ----------------------------------------------------------------------------
>
> Key: KAFKA-16032
> URL: https://issues.apache.org/jira/browse/KAFKA-16032
> Project: Kafka
> Issue Type: Bug
> Components: clients, consumer
> Reporter: Lianet Magrans
> Assignee: Lianet Magrans
> Priority: Major
> Labels: kip-848-client-support
> Fix For: 3.8.0
>
>
> OffsetFetch and OffsetCommit handle errors separately. There are 2 issues to
> review around this:
> - The logic is duplicated for some errors that are treated similarly (ex.
> NOT_COORDINATOR)
> - Some errors are not handled similarly in both requests (ex.
> COORDINATOR_NOT_AVAILABLE handled and retried for OffsetCommit but not
> OffsetFetch). Note that the specific errors handled by each request were kept
> the same as in the legacy ConsumerCoordinator but this should be reviewed, in
> an attempt to handle the same errors, in the same way, whenever possible.
> Note that the legacy approach handles expected errors for each path (FETCH
> and COMMIT), retrying on those when needed, but does not retry on unexpected
> retriable errors.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
