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 storageClasses 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
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.
... snip ...
- added by MDK on 2/16/2023
- reviewed by DK / MAT on 2/16/2023