dynamic-nfs-provisioner

dynamic-nfs-provisioner copied to clipboard

Describe the problem/challenge you have I've deployed the provisioner into my cluster and created an NFS share. I've edited my deployment to provide 2 replicas of the provisioner for maximum resilience. When I then provision an NFS share, the replica set (created by the deployment) to manage the share only has 1 instance which causes two problems for me:

• single point of failure
• hampers migration and node auto-scaling

Describe the solution you'd like I appreciate there may be really good technical reasons (kernel NFS) as to why it isn't possible, but I may as well ask, would it be possible to create a replica set that matches the number of provisioned replicas? e.g.: If I have configured 2 provisioners, each replica set created by the provisions would have 2 replicas.

Anything else you would like to add: N/A

Environment:

• OpenEBS version (use kubectl get po -n openebs --show-labels): NAME READY STATUS RESTARTS AGE LABELS nfs-pvc-a976b2ce-c442-4438-a84b-2ac480ab90f1-785759f588-hk5xj 1/1 Running 0 17h openebs.io/nfs-server=nfs-pvc-a976b2ce-c442-4438-a84b-2ac480ab90f1,pod-template-hash=785759f588 openebs-nfs-provisioner-7b4c9b87d9-w9f7z 1/1 Running 1 18h name=openebs-nfs-provisioner,openebs.io/component-name=openebs-nfs-provisioner,openebs.io/version=dev,pod-template-hash=7b4c9b87d9 openebs-nfs-provisioner-7b4c9b87d9-x8wkc 1/1 Running 0 18h name=openebs-nfs-provisioner,openebs.io/component-name=openebs-nfs-provisioner,openebs.io/version=dev,pod-template-hash=7b4c9b87d9
• Kubernetes version (use kubectl version): Client Version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.5", GitCommit:"6b1d87acf3c8253c123756b9e61dac642678305f", GitTreeState:"clean", BuildDate:"2021-03-18T01:10:43Z", GoVersion:"go1.15.8", Compiler:"gc", Platform:"linux/amd64"} Server Version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.2", GitCommit:"faecb196815e248d3ecfb03c680a4507229c2a56", GitTreeState:"clean", BuildDate:"2021-01-13T13:20:00Z", GoVersion:"go1.15.5", Compiler:"gc", Platform:"linux/amd64"}
• Cloud provider or hardware configuration: Digital Ocean
• OS (e.g: cat /etc/os-release): (Managed Kubernetes: I think the worker nodes are ubuntu 20.04, but not sure)
• kernel (e.g: uname -a): (see above)
• others:

I have run into this same limitation, where nfs file access fails across the cluster if the single node running nfs-pvc fails. As @garypen observes, this is a single point of failure.

I see this is listed as 'To do' in the roadmap. Can you say if solving this is likely to be scheduled any time soon?

Thank you for your valuable open source contributions.