Before You Begin

To access and mount the storage volumes exported by the Lightbits clusters, each of the Kubernetes cluster nodes must already have installed:

• The appropriate Linux kernel.
• The NVMe/TCP transport kernel module.

The Discovery-Client service should be deployed if we want to run it on the host and not inside the lb-csi-node POD.

To ensure smooth access to the storage volumes exported by the Lightbits clusters, the Kubernetes cluster nodes should be configured to automatically load the NVMe/TCP kernel module on Kubernetes node restart.

Before deploying the Lightbits CSI plugin, you should verify that:

• The Lightbits storage cluster is in a "Healthy" state according to its management API service or CLI.
• It is possible to manually mount volumes exported over NVMe/TCP by each of the appropriate Lightbits storage cluster servers on each of the Kubernetes cluster nodes.
• It is possible to manually access the Lightbits management API service instances running on each of the appropriate Lightbits storage cluster servers from each of the Kubernetes cluster nodes; e.g., using curl.
• The Discovery-Client service is started and enabled on each Kubernetes worker node, in case we choose to run it on the host.

Security systems like network firewalls and Linux Mandatory Access Control (MAC) systems (e.g., SELinux, AppArmor) running on the Kubernetes cluster nodes or the Lightbits cluster servers can interfere with proper operation of the Lightbits CSI plugin. You should disable these security systems or configure them to allow the CSI plugin to carry out the requisite actions. For more details, see the Lightbits CSI Plugin Release Notes appropriate to the version of the CSI plugin you are deploying.