This message was deleted.

What version of NV does it try to deploy? Anything in your cluster events? are you deploying on RKE1 cluster from same Rancher UI -> Apps ?

do you have longhorn deployed to that cluster?

I do have longhorn on the cluster, and that's where it allocated the PVC. Looks like the volume is created, but the pods don't ever get past the init state. This is the latest neuvector available app I had (104.0.4+up2.8.4), and neuvector itself is 5.4.2

I would delete it from the app catalog and use the chart itself. the catalog is usually a few months behind. did you

kubectl describe

the pod to see whats up?

I'm looking now. It successfully assigned the pod to a node, but it's sitting at an incomplete status of init prime-config-container

are all the longhorn pods ok? You could modify the app to NOT use persistent storage.

right, and that's what I had been doing. Do you know what gets lost if I don't have persistent storage with neuvector?

all the settings and data.

So, I'd obviously rather have it than not.

makes sense. but NV as never run yet then you can test it.

I think I figured out how to get this working if anyone is interested. It has to do most likely with the default settings of the storageclass as to whether or not the PVC will be created in RWX mode or not. I worked around that issue by pre-creating the PVC neuvector-data and making sure it was set to Many-node read-write. Then instead of checking the box in the Ranch App for a PVC, I edited the yaml and set these options:

pvc:
    accessModes:
      - ReadWriteMany
    capacity: 1Gi
    enabled: true
    existingClaim: neuvector-data
    storageClass: longhorn

This then seemed to deploy ok on RKE1 and RKE2 clusters.