Peter created KAFKA-16384:
-----------------------------
Summary: KRaft controller number recommendation
Key: KAFKA-16384
URL: https://issues.apache.org/jira/browse/KAFKA-16384
Project: Kafka
Issue Type: Bug
Components: controller, kraft
Reporter: Peter
There seems to be some conflicting information about how many controllers
should be used for a KRaft cluster. The first section listed mentions 3 or 5
controllers may be used, but the second section mentions no more than 3 should
be used at the moment.
https://kafka.apache.org/documentation/#kraft_voter
> A Kafka admin will typically select 3 or 5 servers for this role, depending
> on factors like cost and the number of concurrent failures your system should
> withstand without availability impact. A majority of the controllers must be
> alive in order to maintain availability. With 3 controllers, the cluster can
> tolerate 1 controller failure; with 5 controllers, the cluster can tolerate 2
> controller failures.
https://kafka.apache.org/documentation/#kraft_deployment
> For redundancy, a Kafka cluster should use 3 controllers. More than 3
> controllers is not recommended in critical environments. In the rare case of
> a partial network failure it is possible for the cluster metadata quorum to
> become unavailable. This limitation will be addressed in a future release of
> Kafka.
Is 3 still the recommended number and is there more information on what the
network issues are that could cause issues when using 5 controllers?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
