[
https://issues.apache.org/jira/browse/KAFKA-16496?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17860804#comment-17860804
]
Ksolves commented on KAFKA-16496:
---------------------------------
Currently, the receive.buffer.bytes configuration has different default values
across various components:
* connectconfigs: 32k
* adminclientconfigs: 64k
* producerconfigs: 64k
* consumerconfigs: 64k
The suggested changes to update default value of receive.buffer.bytes to -1
instead of 64K needs to be done in all the components?
> Make the default of receive.buffer.bytes -1 instead of 64k
> ----------------------------------------------------------
>
> Key: KAFKA-16496
> URL: https://issues.apache.org/jira/browse/KAFKA-16496
> Project: Kafka
> Issue Type: Improvement
> Reporter: Hechao Li
> Priority: Major
>
> Currently, kafka has a default config receive.buffer.bytes=64k, which
> translates to SO_RCVBUF = 65536. As a result, the Linux kernel can't auto
> tune the receive window. This will cripple performance in many scenarios,
> especially for internet traffic.
> In our environment, we have seen a timeout after [a kernel
> patch|https://lore.kernel.org/netdev/20230717152917.751987-1-eduma...@google.com/T/]
> that makes the initial receive window 25% instead of 50% of SO_RCVBUF,
> together with the default request.timeout.ms=30000, it causes an
> application-level timeout.
> See also https://lore.kernel.org/all/20240409164355.1721078-1-...@netflix.com/
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
