This message was deleted.

If you have a 3 node cluster, why set the replicas to 6 (if I understood correctly)? Also, does your master node have the worker role too? Longhorn will not install on nodes that have a NoSchedule taints. You can see the actual replica count if you click on any volume in the Longhorn UI - Volumes. And you can see the active nodes within the nodes tab in the Lonhorn UI.

@dazzling-france-77424 I've verified both worker nodes. Longhorn instance manager along with few more longhorn pods are running. Not sure, why it's in degraded mode Pods are up n running with PVCs attached to it

Have you checked on one of the Volume the number of replicas within the Longhorn UI?

@dazzling-france-77424 Thanks a lot From the longhorn UI console, I just updated replica count to 2 for each volume and it's healthy now. For my Knowledge I'm asking, if we have 6 worker nodes then the replica should be 6 and like if 3 nodes then replicas should be 3? I'm I right?

It depends the amount of redundancy you want and storage space you have... I don't think it's super useful to have that much replica count unless you have strict redundancy rules or requirements, 3 should be enough (you basically have 3 exact copies of the volume). Data locality might become a "problem" if it is set and you have only 3/6 replicas - then the pods will try to run on a node that has a replica for this specific volume, so if the feature is enabled, you might end up with warnings about data locality. It's really something to think about when you have more than 3 nodes, example, you have 6 and 3 runs on a specific physical server and another 3 runs on another one, then setting replica to 4 will ensure that there is at least 1 replica on each physical server - so if 1 physical server fails, you know your data is still there and safe. Hope that makes sense?