rondagostino commented on a change in pull request #11503:
URL: https://github.com/apache/kafka/pull/11503#discussion_r751651263
##########
File path: core/src/main/scala/kafka/server/KafkaConfig.scala
##########
@@ -2007,12 +2008,47 @@ class KafkaConfig private(doLog: Boolean, val props:
java.util.Map[_, _], dynami
"offsets.commit.required.acks must be greater or equal -1 and less or
equal to offsets.topic.replication.factor")
require(BrokerCompressionCodec.isValid(compressionType), "compression.type
: " + compressionType + " is not valid." +
" Valid options are " +
BrokerCompressionCodec.brokerCompressionOptions.mkString(","))
- require(!processRoles.contains(ControllerRole) ||
controllerListeners.nonEmpty,
- s"${KafkaConfig.ControllerListenerNamesProp} cannot be empty if the
server has the controller role")
+ if (usesSelfManagedQuorum) {
+ require(controlPlaneListenerName.isEmpty,
+ s"${KafkaConfig.ControlPlaneListenerNameProp} is not supported in
KRaft mode.")
+ if (processRoles.contains(ControllerRole)) {
+ // has controller role (and optionally broker role as well)
+ // controller.listener.names must be non-empty
+ // every one must appear in listeners
+ // each port appearing in controller.quorum.voters must match the port
in exactly one controller listener
+ require(controllerListeners.nonEmpty,
+ s"${KafkaConfig.ControllerListenerNamesProp} must contain at least
one value appearing in the '${KafkaConfig.ListenersProp}' configuration when
running the KRaft controller role")
+ val listenerNameValues = listeners.map(_.listenerName.value).toSet
+ require(controllerListenerNames.forall(cln =>
listenerNameValues.contains(cln)),
+ s"${KafkaConfig.ControllerListenerNamesProp} must only contain
values appearing in the '${KafkaConfig.ListenersProp}' configuration when
running the KRaft controller role")
+ RaftConfig.parseVoterConnections(quorumVoters).asScala.foreach { case
(nodeId, addressSpec) =>
Review comment:
> We should only check the port of this voter/controller
I coded it that way first, but upon further thought I came to the conclusion
that we should check every voter -- not just this one. Based on my comment
above, I believe this should actually require the port to match the first
controller listener's port, and that further adds weight to checking every node.
--
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-unsubscr...@kafka.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
