Longhorn distributed storage

#1

I have 3 nodes within Rancher that I would like to show in the Longhorn dashboard.

When I do install longhorn from the catalog apps to each cluster, it only shows 1 node in the longhorn dashboard. The odd thing is that it seems to show the storage from another node. For example, if I access node-3 -> catalog apps -> index.html, I will land on a dashboard where I see information regarding node-1’s storage.

I would like to configure it to show all nodes in the dashboard. However, I can’t seem to find any instructions on this. Can you help with some guidance?

I would like the end result to be something similar to this: https://www.cnrancher.com/blog/2018/2018-08-23-longhorn-v0.3.0-release/

Notes:

  • I am not running Rancher in HA (yet)
  • I am using the default Canal networking with Project Network Isolation disabled.
#2

@mattsdevop

The catalog apps are not per node, it’s per cluster/project. So I am not sure what do you mean if you access node-3 -> catalog apps.

Longhorn will only use the worker nodes of Kubernetes cluster. Can you show me the result of kubectl get nodes as well as kubectl -n longhorn-system get pod?

1 Like
#3

Thank you for the reply @yasker Apologies for using the wrong terminology.
Access cluster-3 -> catalog apps -> index.html. And, it will show the information for the server connected to cluster-1.
I do not currently have Rancher in the same state. I will reconfigure and run the commands requested.

#4
> kubectl get nodes
NAME              STATUS    ROLES                      AGE       VERSION
cluster-01   Ready     controlplane,etcd,worker   2h        v1.11.6
> kubectl -n longhorn-system get pod
NAME                                        READY     STATUS    RESTARTS   AGE
csi-attacher-7fc5746899-frhn6               1/1       Running   0          2m
csi-attacher-7fc5746899-qtqpz               1/1       Running   0          2m
csi-attacher-7fc5746899-wrlr8               1/1       Running   0          2m
csi-provisioner-74b6dfb5bb-5v97w            1/1       Running   0          2m
csi-provisioner-74b6dfb5bb-cbw6r            1/1       Running   0          2m
csi-provisioner-74b6dfb5bb-qndrb            1/1       Running   0          2m
engine-image-ei-6e2b0e32-fqrd5              1/1       Running   0          3m
longhorn-csi-plugin-qgzff                   2/2       Running   0          2m
longhorn-driver-deployer-587d4db779-p8jgr   1/1       Running   0          3m
longhorn-manager-tvtx2                      1/1       Running   0          3m
longhorn-ui-5486dcd785-qbsdx                1/1       Running   0          3m

Does Rancher need to be run within a kubernetes cluster for this to work as I imagined it? So that kubectl get nodes returns all of my cluster nodes? Just trying to figure out what I’m missing here. Thanks!

Edit: This time accessing the index.html from cluster-2 -> catalog apps -> index.html takes me to the dashboard. But, it shows cluster-3 information.

#5

I believe I’ve found my mistake. I was misinformed about the initial configuration. We have multiple clusters, and not nodes. I’ve reconfigured and will now test. Thank you for the direction @yasker