You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "Ritabear (via GitHub)" <gi...@apache.org> on 2026/10/27 11:55:54 UTC
[PR] KAFKA-17993: Fall back to wildcard when the throttled replica list is too long [kafka]
Ritabear opened a new pull request, #23980:
URL: https://github.com/apache/kafka/pull/23980
When many partitions of a topic are reassigned with a throttle, the throttled
replica list can exceed the maximum config value size (32767 bytes). The
request then fails to serialize and kafka-reassign-partitions.sh hangs.
Splitting the value with APPEND does not help, as the controller stores it as a
single ConfigRecord. This patch uses the wildcard `*` for such a topic instead.
Only out-of-sync replicas are actually throttled, so the impact on partitions
that are not being reassigned is limited.
Generated-by: Claude Opus 5.5
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: jira-unsubscribe@kafka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] KAFKA-17993: Fall back to wildcard when the throttled replica list is too long [kafka]
Posted by "Ritabear (via GitHub)" <gi...@apache.org>.
Ritabear commented on PR #23980:
URL: https://github.com/apache/kafka/pull/23980#issuecomment-6441266082
Added an integration test that sets a ~38KB throttled replica list. Without the
fix, the admin client thread dies with "'value' field is too long to be
serialized" and the call never completes.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: jira-unsubscribe@kafka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] KAFKA-17993: Fall back to wildcard when the throttled replica list is too long [kafka]
Posted by "Ritabear (via GitHub)" <gi...@apache.org>.
Ritabear commented on PR #23980:
URL: https://github.com/apache/kafka/pull/23980#issuecomment-6443889907
@chia7712 could you take a look when you have time? This follows the wildcard fallback you suggested on the JIRA.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: jira-unsubscribe@kafka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org