Citrix Node Controller Version 2.2.2

Version 2.2.2

What's New

• A mandatory environment variable CNI_TYPE is added for specifying the container network interface (CNI) used in the Kubernetes cluster.
• Weave CNI is now supported.
• Now, kube-cnc-router pods and ConfigMap are moved from the kube-system namespace to the CNC pod namespace.
• The term routervxlan0 for node VTEP interface is now changed to cncvxlan<hash_of_namespace>
• If required, the Citrix node controller ConfigMap data section can be used to specify tolerations for taints.
• ConfigMap for the Citrix node controller is handled only if the ConfigMap is configured in the CNC pod namespace.
• Multiple CNCs can be configured in different namespaces (for different ADCs) while keeping the overlay NETWORK as non-conflicting.
• A troubleshooting guide is now added with information for resolving the common issues while deploying.

Fixed Issues

• Now, nodes do not require the NodeId- label.
• Now, worker node addition for the Canal CNI is handled properly.
• Node reboot is now properly handled for the Flannel CNI, if no other pods are present.
• Some of the unnecessary pod privileges for kube-router-pods are removed.

CNC Version 2.2.1

Version 2.2.1

What's new

• An optional environment variable DSR_IP_RANGE is added for DSR configurations on the Citrix ADC.
• Now, the Citrix ADC is configured with static bridge table configurations instead of static ARP configurations.
• The new Calico CNI modes are supported.
• The term Kube-chorus-router is now changed to kube-cnc-router.

Fixed Issues

• When you delete a Citrix Kubernetes Node Controller ConfigMap, the ADC configuration is automatically removed from the Citrix ADC and Kubernetes deployments for kube-cnc-router pods, service account, and cluster role are removed.

• Now, worker node restart events are handled properly.

Version 2.1.0

Version 2.1.0

This release note provides information about the fixed issues and known issues in the Citrix Kubernetes Node Controller version 2.1.0.

Fixed issues

• The Citrix Kubernetes Node Controller was crashing when Calico is used as the Container Network Interface (CNI). Now, this issue is fixed.

Known issues

• When you delete a Citrix Kubernetes Node Controller pod, node extender pods and ConfigMap are not deleted.

• When you delete a Citrix Kubernetes Node Controller pod configuration, the configuration is not automatically removed from the Citrix ADC and you have to manually remove it.

• Node restart events are currently not handled properly. If a node is shut down, then the corresponding route configuration is not removed from the Citrix ADC. Similarly, other node events are also not properly handled. Hence, it is recommended to avoid node shutdowns.