23. Release Notes¶

23.1. Robin Cloud Native Platform v5.3.9¶

The Robin Cloud Native Platform (CNP) v5.3.9 has new features, improvements, bug fixes, and known issues.

Release Date: 06 October 2021

23.1.1. Infrastructure Versions¶

The following software applications are included in this CNP release.

Software Application	Version
Kubernetes	1.21.5
Docker	19.03.9
Prometheus	2.16.0
Node-exporter	1.1.2
Calico	3.12.3
HA-Proxy	1.5.18
PostgreSQL	9.6.11
Grafana	6.5.3

23.1.2. Upgrade Paths¶

The following are the upgrade paths for Robin CNP v5.3.9:

Robin v5.3.3-115 (HF4) to Robin v5.3.9-286 (GA)
Robin v5.3.5-159 (HF2) to Robin v5.3.9-286 (GA)
Robin v5.3.5-207 (HF3) to Robin v5.3.9-286 (GA)
Robin v5.3.7-120 (HF1) to Robin v5.3.9-286 (GA)

23.1.3. New Features¶

23.1.3.1. Auto Disk Rebalance¶

Robin now supports the Auto Disk Rebalance feature. The Auto Disk Rebalance feature manages the storage space of all disks in the cluster automatically when the disk reaches a certain watermark threshold.

By default, the Auto Disk Rebalance feature is enabled.

When a disk reaches a high watermark, the disk rebalance job automatically starts to move the volumes from one disk to another disk. The Auto Disk Rebalance feature is designed in such a way that it always selects the disk which has more free space.

23.1.3.2. Support for Vault Integration as KMS¶

Robin supports the integration of HashiCorp Vault as a key management service (KMS) with Robin Cloud Native Platform (CNP). You can integrate Vault as part of Robin CNP installation only. You cannot integrate Vault with Robin CNP after installation and with the existing clusters.

23.1.3.3. Network Policies Support¶

Robin supports the Kubernetes NetworkPolicies for Robin bundles and Helm applications based on the Calico network. Kubernetes NetworkPolicies allow you as an administrator to enforce rules on using network traffic. You can use the network policies in Kubernetes to specify how groups of Pods are allowed to communicate with each other and with external network endpoints.

Robin integration with Kubernetes NetworkPolicies helps in achieving Robin user, tenant, or namespace level network ingress/egress traffic isolation for Calico network-based workloads.

23.1.3.4. Persistent Data Volumes¶

Robin CNP now supports Persistent Data Volumes (PDV). A PDV in Robin is a volume that persists even after an application is deleted from the cluster. A PDV is derived from the persistent storage volume (PV).

A PDV provides tenant users a private volume just like a personal home folder in a file server environment. Only the owner of the PDV can access the PDV.

There are two types of PDVs:

Private PDV - Tenant users can create a private PDV. Tenant users can only see the PDVs owned by them.

Public PDV - Cluster or tenant admins can only create a Public PDV and the admins can share it with multiple tenants/users.

23.1.4. Improvements¶

23.1.4.1. HA-Proxy and Keepalived Run as a Docker Container¶

Starting with Robin 5.3.9, the HA-Proxy and Keepalived run along with robink8s_monitor in the Docker container.

Running these components in a Docker container avoids impacting critical and high available cluster services.

Prior to Robin CNP 5.3.9, HA-Proxy and Keepalived run on the host as system services.

23.1.4.2. Kafka Notifications Display Cluster UUID and Master Hostname¶

Starting with Robin v5.3.9, Kafka notifications include the cluster UUID and master hostname.

23.1.4.3. Robin CNP Installer with Cluster UUID Option¶

Robin CNP Installer now accepts a cluster UUID option using –cluster-uuid during the first master install. This can be used by the caller to stamp a cluster UUID of their choice on the CNP cluster.

23.1.4.4. Support to Add gzipped qcow2 Images as KVM Images¶

Robin CNP now supports adding gzipped qcow2 images as KVM images. Before uploading the qcow2 image to Robin CNP, you can gzip the file and register it with Robin via the robin image add command, detailed here.

23.1.4.5. Support for Automatically Mount Volumes for KVM Applications¶

Robin CNP v5.3.9 supports automatically mounting volumes when creating KVM-based applications.

23.1.4.6. View Helm App Specifications using API¶

Starting with Robin CNP 5.3.9, you can use the same API to view Helm app details that is used for Robin Bundles.

23.1.4.7. Integrated Key Wrap Algorithm¶

Robin CNP 5.3.9 is now integrated with the Advanced Encryption Standard key wrap algorithm adhering to the RFC3394.

23.1.4.8. Support to Configure Process IDs Limit¶

Robin CNP 5.3.9 supports configuring process IDs limit when installing or upgrading Robin CNP. The default configuration is 1000.

23.1.4.9. Support to Configure Pod Limit Range¶

Robin CNP 5.3.9 supports configuring the Pod limit range for the bundle and non-bundle applications. This rejects Pods without limits based on a config variable. enable_limit_range is the flag. If this is enabled, Pods that do not have limits set, will be rejected.

23.1.4.10. New Docker Version¶

Robin CNP 5.3.9 now supports Docker Version 19.03.9.

23.1.4.11. New Node-Exporter Version¶

Robin CNP version 5.3.9 now supports node-exporter version 1.1.2.

23.1.5. Fixed Issues¶

Reference ID	Description
PP-23126	The issue of VIPs frequently flapping across master nodes and leading to health check time out due to Keepalived packages upgrading to a greater version than supported by Robin is fixed.
PP-22121	The issue of a pod being stuck in the ContainerCreating state, when its definition contains a secondary IP with a gateway has been fixed.
PP-23942	The issue of the PostgreSQL slave instance not coming up after upgrading the Robin CNP is fixed.
PP-23739	The issue of allocated hugepages not being displayed as part of the output of the `robin host info` command has been fixed.
PP-23477	The issue cluster reporting containers stuck in ContainerCreating state due to NFS-Server Pod failover and Exports failing jobs, after upgrading to Robin CNP version is fixed.
PP-23954	The issue of IO Manager failed to come up when upgrading a cluster to 5.3.5 HF3 is fixed.

23.1.6. Known Issues¶

Reference ID	Description
PP-22781	Symptom After removing a taint on a master node, GPUs are not detected automatically. GPUs doesn’t get detected automatically after removing taint on master node. Workaround You need to run the `robin host probe --rediscover --all --wait` command for the GPUs to be detected on the primary master node.
PP-22626	Symptom If NVIDIA GPU drivers are already installed on the setup, operator deployment might fail. Workaround Run the following steps on the node where the issue is encountered: yum remove nvidia-driver-latest-dkms yum remove nvidia-container-toolkit Reboot the node
PP-24051	Symptom After upgrading to Robin CNP 5.3.9 with `toplogy_manager_policy=best-effort` and vDUs, the vDUs might not come back online because vfs driver vfio-pci not released during app drain. Workaround Use the vfio-pci driver for sriov instead of iavf.
PP-23707	Symptom When you run some concurrent app creations and deletions operations of the same bundle, the app creation might fail due to an internal race condition. Workaround You can restart the vnode or try to create the application again.

23.1.7. Technical Support¶

Contact Robin Technical support for any assistance.