0.4.0

New features/Bug fixes

• Multi-cluster provisioning support enables provisioning and distribution of the service instances into multiple clusters. Architecture Docs
• Round-robin Scheduler Support
• Public helm chart repository for interoperator.

Supported K8S Version

• 1.13.x
• 1.14.x
• 1.15.x

The release is not yet tested against 1.16.x.

How to deploy Interoperator

To add service fabrik interoperator helm chart repo

helm repo add sf-charts https://cloudfoundry-incubator.github.io/service-fabrik-broker/helm-charts
helm repo update

Deploy SF Interoperator using helm

helm install --set cluster.host=sf.ingress.< clusterdomain > --name interoperator --namespace interoperator --version 0.4.0 sf-charts/interoperator

NOTE: cluster.host should be within the 63 character limit.

The default scheduler which gets deployed by the helm chart supports only one cluster with name "1". The scheduler type can be configured using --set interoperator.config.schedulerType=<type> parameter to the helm install/upgrade command. The current supported scheduler tyes are default and round-robin.

Deploy SFClusters, SFServices and SFPlans and Register with Interoperator

Please create sfcluster CRs and add reference to kubeconfig.

For multi-cluster support, all corresponding sfcluster CRs need to be created and their kubeconfig needs to be supplied in the corresponding secret.

Please note that sfservice and sfplans need to be deployed in the same namespace where SF is deployed (default is interoperator).

Upgrade from the earlier releases(special handling, downtime if any)

To add service fabrik interoperator helm chart repo if not already added

# Assuming the repo name is chosen as sf-charts 
helm repo add sf-charts https://cloudfoundry-incubator.github.io/service-fabrik-broker/helm-charts
helm repo update

To update to 0.4.0 version the following steps are to be done in order to upgrade using helm. This would involve few seconds of downtime as once the ClusterRole is deleted the existing deployment stops working and there is downtime till helm upgrade is completed.

kubectl delete ClusterRole interoperator-interoperator-manager-role

# Assuming interoperator is currently deployed in interoperator namespace
kubectl -n interoperator delete ConfigMap interoperator-config

Helm upgrade should take care of upgrading to the latest release.

# Assuming current helm release name is interoperator 
helm upgrade --set cluster.host=sf.ingress.< clusterdomain > --namespace interoperator --version 0.4.0 interoperator sf-charts/interoperator --force --recreate-pods