0.5.0-rc2

Changes since 0.5.0-rc1

• Avoid patching ServiceAccounts to avoid continual token creation #641

0.5.0-rc1

Main features

• The Cluster Operator now manages RBAC resource for managed resources:
  • ServiceAccount and ClusterRoleBindings for Kafka pods
  • ServiceAccount and RoleBindings for the Topic Operator pods
• Renaming of Kubernetes services (Backwards incompatible!)
  • Kubernetes services for Kafka, Kafka Connect and Zookeeper have been renamed to better correspond to their purpose
  • xxx-kafka -> xxx-kafka-bootstrap
  • xxx-kafka-headless -> xxx-kafka-brokers
  • xxx-zookeeper -> xxx-zookeeper-client
  • xxx-zookeeper-headless -> xxx-zookeeper-nodes
  • xxx-connect -> xxx-connect-api
• Cluster Operator moving to Custom Resources instead of Config Maps
• TLS support has been added to Kafka, Zookeeper and Topic Operator. The following channels are now encrypted:
  • Zookeeper cluster communication
  • Kafka cluster commbunication
  • Communication between Kafka and Zookeeper
  • Communication between Topic Operator and Kafka / Zookeeper
• Logging configuration for Kafka, Kafka Connect and Zookeeper
• Add support for Pod Affinity and Anti-Affinity
• Add support for Tolerations
• Configuring different JVM options
• Support for broker rack in Kafka

---

All changes merged for this release should be listed under the 0.5.0 milestone.

0.4.0

Changes since 0.3.1:

• Better configurability of Kafka, Kafka Connect, Zookeeper
• Support for Kubernetes request and limits
• Support for JVM memory configuration of all components
• Controllers renamed to operators
• Improved log verbosity of Cluster Operator
• Update to Kafka 1.1.0

List of all issues can be found here.

This release has some backwards incompatible changes:

• Cluster Controller and Topic Controller were renamed to Cluster Operator and Topic Operator
  • Roles, role bindings and service accounts existing from previous deployments can not be reused
• The following fields from Kafka ConfigMap have been replaced with new field kafka-config
  • KAFKA_DEFAULT_REPLICATION_FACTOR
  • KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR
  • KAFKA_TRANSACTION_STATE_LOG_REPLICATION_FACTOR
• The following fields from Kafka Connect ConfigMap have been replaced with new field connect-config
  • KAFKA_CONNECT_BOOTSTRAP_SERVERS
  • KAFKA_CONNECT_GROUP_ID
  • KAFKA_CONNECT_KEY_CONVERTER
  • KAFKA_CONNECT_KEY_CONVERTER_SCHEMAS_ENABLE
  • KAFKA_CONNECT_VALUE_CONVERTER
  • KAFKA_CONNECT_VALUE_CONVERTER_SCHEMAS_ENABLE
  • KAFKA_CONNECT_CONFIG_STORAGE_REPLICATION_FACTOR
  • KAFKA_CONNECT_OFFSET_STORAGE_REPLICATION_FACTOR
  • KAFKA_CONNECT_STATUS_STORAGE_REPLICATION_FACTOR

0.4.0-rc1

Changes since 0.3.1:

• Better configurability of Kafka, Kafka Connect, Zookeeper
• Support for Kubernetes request and limits
• Support for JVM memory configuration of all components
• Controllers renamed to operators
• Improved log verbosity of Cluster Operator
• Update to Kafka 1.1.0

List of all issues can be found here

0.3.1

Changes since 0.3.0:

• Use correct default image for Kafka Connect S2I deployments #416

0.3.0

Main changes since 0.2.0:

• Improved Cluster controller resiliency
• DNS names used for advertised listeners
• Improved test coverage
• Improved documentation
• Configuring default Docker images in Cluster Controller

0.3.0-rc2

RC2 fixes the following problems from RC1:

• Constant rolling update of Kafka and Zookeeper with persistent storage on Kubernetes (PR #384).
• Inconsistencies when user attempts to change the storage type for running cluster (PR #386)

0.3.0-rc1

Changes since 0.2.0:

• Improved resiliency
• DNS names used for advertised listeners
• Improved test coverage
• Improved documentation
• Configuring default Docker images in Cluster Controller

0.2.0

Main changes:

• New operator concept with Cluster Controller and Topic Controller
  • Custer Controller is responsible for deployment of Kafka and Kafka Connect clusters
  • Topic Controller is managing Kafka Topics
• Improved documentation
• Improved examples
• Docker images based on CentOS 7
• Improved test coverage
• Update to Kafka 1.0.1

Known issues:

• #264

0.2.0-rc3

Changes since rc2:

• Release lock when periodic reconcile is freed after ongoing deletion #296
• Persistent Volume name too long #294
• Update to Kafka 1.0.1 #242