Configuring Separate Networks for DRBD's Replication When using the LINSTOR Operator v1 for Kubernetes

It’s considered a best practice to use a separate dedicated network for DRBD® replication. This article describes how to do this when using the LINSTOR® Operator (v1.x) to deploy LINSTOR into Kubernetes.

The LINSTOR Operator v1 does not support configuring a separate replication network during its initial deployment of LINSTOR. However, once you’ve deployed LINSTOR, and before you define your LINSTOR storage classes in Kubernetes, you can use the following commands and configurations to separate your replication traffic.

Assume you have three nodes named linbit1, linbit2, and linbit3, with dedicated network interfaces for DRBD’s replication configured with the IP addresses, 10.10.10.101, 10.10.10.102, and 10.10.10.103, respectively. Also assume that you’ve deployed LINSTOR by using the LINSTOR Operator v1.x into your Kubernetes cluster with the CNI configured to use some other network.

Start a shell session on the LINSTOR Controller pod and add an additional interface to each node in your LINSTOR cluster. This example uses the name rep_nic, but this can be whatever you’d like.

kubectl exec -it deployments/linstor-op-cs-controller -- bash
linstor node interface create linbit1 rep_nic 10.10.10.101
linstor node interface create linbit2 rep_nic 10.10.10.102
linstor node interface create linbit3 rep_nic 10.10.10.103
exit

With the dedicated network interfaces configured in LINSTOR, you can create the LINSTOR storageClass definitions with the PrefNic parameter, instructing LINSTOR to use the dedicated network instead of the default network.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: linstor-storage-pool-example
provisioner: linstor.csi.linbit.com
parameters:
  ... snip ...
  property.linstor.csi.linbit.com/PrefNic: rep_nic

Written by MDK on 2/16/2023

Reviewed by DK / MAT on 2/16/2023