@ancient-wolf-19325 I'm trying to mount an NFS filestore in one of my jobs, but it doesn't seem to work:

# mount -o nolock <ip>:/blah /mnt/blah
mount: /mnt/balh: cannot mount <ip>:/blah read-only.

I've checked connectivity (works) and I've also been able to mount it in a different VM on the same VPC. I've even been able to mount it on the flyte node instance. Is there something in the kubernetes config that is preventing this?

The issue might be related to Kubernetes permissions or the configuration of the NFS mount in the pod specification. According to Flyte GitHub Issue #2754, you can use a

V1NFSVolumeSource

in the pod spec to mount an NFS. However, permission issues can arise if the security context is not set correctly. You might need to adjust the

fsGroup

in the

securityContext

to match the group ID that has access to the NFS. Would you like more detail on how to configure the security context or other potential solutions? Sources: - Flyte GitHub Issue #2754 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@ancient-wolf-19325 more detail pleas yes

To resolve the issue of mounting an NFS filestore in your Kubernetes pod, you need to ensure that the security context is correctly configured to allow access to the NFS. Here's a step-by-step breakdown: 1. Check NFS Permissions: Ensure that the NFS server allows the Kubernetes nodes to access the NFS share. This includes verifying that the IP range of the nodes is permitted and that the correct permissions are set on the NFS server. 2. Configure Pod Spec with NFS Volume: Use the

V1NFSVolumeSource

in your pod specification to define the NFS mount. This involves specifying the server IP and the path to the NFS share. 3. Set Security Context: Adjust the

securityContext

in your pod spec to ensure that the pod has the necessary permissions to access the NFS. Specifically, set the

fsGroup

to the group ID that has access to the NFS. This ensures that the pod can read and write to the NFS share. Example YAML configuration:

apiVersion: v1
   kind: Pod
   metadata:
     name: nfs-client
   spec:
     containers:
     - name: app
       image: your-image
       volumeMounts:
       - name: nfs-volume
         mountPath: /mnt/blah
     volumes:
     - name: nfs-volume
       nfs:
         server: <ip>
         path: /blah
     securityContext:
       fsGroup: 1000  # Replace with the correct group ID

4. Verify Kubernetes Configuration: Ensure that your Kubernetes cluster is configured to allow NFS mounts. This includes checking any network policies or security settings that might block NFS traffic. 5. Test the Configuration: Deploy the pod and check the logs to ensure that the NFS mount is successful. If there are errors, they might provide additional clues about what needs to be adjusted. By following these steps, you should be able to successfully mount an NFS filestore in your Kubernetes pod. If you continue to encounter issues, consider checking the Kubernetes and NFS server logs for more detailed error messages. Sources: - Flyte GitHub Issue #2754 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.