I'm facing some challenges with Longhorn and I'm looking for help. My goal is to set up Longhorn storage in such a way that it:
1. Limits storage writes to three nodes that have disks installed.
2. Ensures all nodes can access Longhorn storage.
3. Is deployed using Helm.
4. Can be configured with a config.yaml file.
I've already labeled and tainted my nodes accordingly, but Longhorn seems to have issues with this setup. If anyone has a sample config.yaml or some insights, that would be really appreciated!
1 Answer
Check out the documentation—I found it quite reliable! I activated the option "Create Default Disk on Labeled Nodes". Make sure to apply this label on the nodes that have disks:
node.longhorn.io/create-default-disk=true
It worked wonders for me! This way, all nodes can use Longhorn storage classes, but disks are only created on the labeled nodes.
Thanks! I'm having trouble with Longhorn placing stuff on my control nodes. Any tips on preventing that? Should I just apply a taint on the control nodes?