Releases: banzaicloud/koperator
Kafka Operator v0.16.0
This release fixes several bug and introduce a couple of new features:
New features:
- Move to crd version v1 and update kubebuilder
- Update Kafka image to 2.7 and update CC image accordingly
- Add more documentation and sample files
- Enable multiple linters
- Allow to use SASL listener types
- Support adding broker annotations to brokers added by 'upScale' alert
- Update istio operator version to 0.9.1
- Allow to add additional env variables to istioingress resource
Bug Fixes:
- Remove ownerref from cluster scoped clusterissuer
- Fix: Labels not getting updated for PVCs during reconciliation
Upgrade Notes:
This release will trigger a rolling update.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.15.1
This release fixes a bug:
Bug Fixes:
- Fix ingress controller behaviour in case of broker downscale
Upgrade Notes:
This release will trigger a rolling update.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.15.0
This release fixes several bug and introduce a couple of new features:
New features:
- Allow to add service account for operator and the authProxy
- Introduce Ginkgo based tests
- Allow to use all dynamic config values as dynamic config no more unnecessary rolling updates
- Introducing AnyCastPort setting
- Switch to golang 1.15
- Add new listenerStatuses status field to the CR
- Implement pod security context for CC and Kafka broker pods
- Make rolling update more reliable by deleting the controller broker only the end of the process
- Add new properties parsing library
- Update Banzai Istio Operator dependency to 0.8.6
- Introduce multiple external listener support to the operator
- Introduce external Listener binding to specific broker
- Allow to add InitContainers and Volumes to the brokers.
For all the changes please check the link provided in the release below
Bug Fixes:
- Fix deadlock which occurs when the controller broker id is overwritten in the CR
Upgrade Notes:
- This release will trigger a rolling update.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.14.0
This release fixes several bug and introduce a couple of new features:
New features:
- Support for multiple external listener
- Support CC version 2.5.23
- Support NodePort external access
- Move all images to ghcr.io
- Update kafka jmx-exporter metrics rules
Bug Fixes:
- Fix ssl cert generation when headless service is not used
Upgrade Notes:
- This release will trigger a rolling update.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.13.1
This release fixes several bugs:
Bug Fixes:
- Fix envoy config generation when used without headless service
- Add urisans field to generated certs
- Improve shutdown script to handle network failure faster
Upgrade Notes:
- This release will trigger a rolling update and will also regenerate all certificates.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.13.0
This release fixes several bug and introduce a couple of new features:
New features:
- Build docker images to Github registry
- Use CC image 2.5.16
- Update jmx exporter to version 0.14 and enable cacheing
- Allow to set log4j config for CC and Kafka
- Allow to not generate cc config hash
- Allow to use external init containers for Kafka Brokers
- Compute CC CPU utilisation dynamically
- Use more precise default config for CC nw_out nw_in metrics and allow to configure it via the user
- Add PodDistruptionBudget support
- Improve logging
Bug Fixes:
- Clean removal of the Broker from Envoy/Istio ingress controllers in case of cluster downscale
- Simplify and fix Vault PKI related code
- Use clusterNamespace in cluster issuer to make it unique
Upgrade Notes:
To upgrade the chart since the helm 3 limitation you have to set a value as well to keep your CRDs.
If this value is not set your CRDs might be deleted.
helm upgrade kafka-operator --set crd.enabled=true --namespace=kafka banzaicloud-stable/kafka-operator
If you are using CRDs installed by kubectl apply then crd.enabled=true can be left out.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
v0.12.4..v0.13.0
This upgrade will trigger a rolling upgrade.
Thanks for the huge help from the community 🍺
Kafka Operator v0.12.4
This release fixes several bugs:
Bug Fixes:
- Bump CC image to version 2.5.11
- Use the official envoy image
- Fix resize PVC alert command
Upgrade Notes:
To upgrade the chart since the helm 3 limitation you have to set a value as well to keep your CRDs.
If this value is not set your CRDs might be deleted.
helm upgrade kafka-operator --set crd.enabled=true --namespace=kafka banzaicloud-stable/kafka-operator
This update will trigger a rolling upgrade.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.12.3
This release fixes several bug and introduce a couple of new features:
New features:
- Do not create serviceAccount if using the default (This behaviour changes since it used kafka-cluster as a default in release v0.12.2)
- Make Helm chart helm3 compatible
- Upgrade jmx version to 0.13.0
- Add DescribeConfig ACL every time when granting an ACL
- Update go version to 1.14
- Update CC to version 2.5.x
- Update CertManager to version 0.15.x and move JKS generation from the operator to the cert-manager
- Add the ability to set annotations to CC pods
- Allow passing loadbalancer IP to the envoy loadbalancer from CR
- Implement Resize PVC command
- Add support for kafka 2.6.0
Bug Fixes:
- Remove finalizers when CR is in deletion state without certificates
- Controller test fixes
- Fix Operator does not restart pod when initcontainer fails with error
- Do not terminate sidecar proxy until broker is running
Upgrade Notes:
- Since ServiceAccount generation changed. Now the operator will use default if not specified otherwise. So if you want to keep using serviceAccount
kafka-clusterplease add that directly to your CR before the version update.
brokerConfigGroups:
default:
serviceAccountName: kafka-cluster
This update will trigger a rolling upgrade.
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.12.2
This release fixes several bugs:
Bug Fixes:
- Add support for new version of Banzai Clouds Istio Operator(Istio 1.6)
- Export java.{lang,nio} broker JMX metrics
- Fix panic when pvc size is updated externally
- Upgrade k8sobjectmatcher to version 1.3.3
This release supports Cert-Manager 0.11
Upgrade Notes:
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺
Kafka Operator v0.12.1
This release fixes several bugs:
Bug Fixes:
- Fix cluster is continuously reconciled if either serviceAnnotation or virtualServiceAnnotation field is changed
- Fix Adjusting storage request in KafkaCluster BrokerConfig does not update PVC
- Upgrade Cruise control to use the new 2.4.x releases (Fixes bug which happens because the old CC does not use the Kafka Admin API for rebalance)
This release supports Cert-Manager 0.11
Upgrade Notes:
Please add the following line to your Cruise-control config section of the KafkaCluster CR.
metric.anomaly.detection.interval.ms=180000
Note: Seamless upgrade only works when your cluster is in ClusterRunning state.
Thanks for the huge help from the community 🍺