Feature Request: fix IEAM agent installation timeout due to PV/PVC not bounding

[yubinqiu1101]

Is your feature request related to a problem? Please describe.

We try to install IEAM agent followed by IEAM installation document
Installation will freeze about 75 seconds and forcely quite .
After checking logs, the agent pod is waiting for PVC/PV binding.
However , the PV is never created
Our tmp solution is create a PV with the name IEAM PVC requested .
apiVersion: v1 kind: PersistentVolume metadata: name: openhorizon-agent-pv spec: storageClassName: "topolvm-provisioner" persistentVolumeReclaimPolicy: Delete claimRef: name: openhorizon-agent-pvc namespace: openhorizon-agent accessModes: - ReadWriteOnce capacity: storage: 10Gi local: path: /root/agent-pv nodeAffinity: required: nodeSelectorTerms: - matchExpressions: - key: node-role.kubernetes.io/control-plane operator: Exists
After the PV get ready in our cluster, we can successfully install the IEAM (HZN) agent
For the uninstallation and re-installation , we need to clean up the pv manually created. Delete the PV in K8S level and delete folder and re-create folder in the node

Here is a feature request that : IEAM installation should make the installation smooth in all linux/k8s environment. Creating custom PV will only make client/customer more complexity

Describe the solution you'd like.

In the IEAM agent installation script , a PVC will be created which is waiting for PV
Sometimes default storageClass may make PV not available to be ready status in some K8S environment.
I will recommand to create a IEAM storageClass which can define how the PV is created and how it is bound with PVC .

Describe alternatives you've considered

1. create custom storageclass

Additional context.

No response

[dlarson04]

hi @yubinqiu1101 - I am not understanding why the PV is not getting created as part of the OH install....

In your environment, are you setting the storage class to topolvm-provisioner ?
If you are setting that, does the PV eventually get created but it just takes > 75 seconds? We have actually changed the default timeout to 300 but there is a variable to set this to whatever you want.

If it is failing, can you do kubectl get events -A so I can see what might be failing.
Finally, what type of k8s cluster are you using and what architecture?