Kubernetes PersistentVolume on local machine, share data - kubernetes

Question

I would like to spin up a Pod on my local machine. Inside the pod is a single container with a .jar file in it. That jar file can take in files, process then, and then output them. I would like to create a PersistentVolume and attach that to the Pod, so the container can accesss the files.
My Dockerfile:
FROM openjdk:11
WORKDIR /usr/local/dat
COPY . .
ENTRYPOINT ["java", "-jar", "./tool/DAT.jar"]
(Please note that the folder used inside the container is /usr/local/dat)
My PersistentVolume.yml file:
apiVersion: v1
kind: PersistentVolume
metadata:
name: dat-volume
spec:
accessModes:
- ReadWriteOnce
capacity:
storage: 150Mi
storageClassName: hostpath
hostPath:
path: /home/zoltanvilaghy/WORK/ctp/shared
My PersistentVolumeClaim.yml file:
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: dat-pvc
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 100Mi
storageClassName: hostpath
volumeName: dat-volume
My Pod.yml file:
apiVersion: v1
kind: Pod
metadata:
name: dat-tool-pod
labels:
name: dat-tool-pod
spec:
containers:
- name: dat-tool
image: dat_docker
imagePullPolicy: Never
args: ["-in", "/usr/local/dat/shared/input/Archive", "-out", "/usr/local/dat/shared/output/Archive2", "-da"]
volumeMounts:
- mountPath: /usr/local/dat/shared
name: dat-volume
restartPolicy: Never
volumes:
- name: dat-volume
persistentVolumeClaim:
claimName: dat-pvc
If all worked well, after attaching the PersistentVolume (and putting the Archive folder inside the shared/input folder), by giving the arguments to the jar file it would be able to process the files and output them to the shared/output folder.
Instead, I get an error saying that the folder cannot be found. Unfortunately, after the error the container exists, so I can't look around inside the container to check the file structure. Can somebody help me identify the problem?
Edit: Output of kubectl get sc, pvc, pv :
NAME PROVISIONER RECLAIMPOLICY VOLUMEBINDINGMODE ALLOWVOLUMEEXPANSION AGE
storageclass.storage.k8s.io/hostpath (default) docker.io/hostpath Delete Immediate false 20d
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
persistentvolumeclaim/dat-pvc Bound dat-volume 150Mi RWO hostpath 4m52s
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
persistentvolume/dat-volume 150Mi RWO Retain Bound default/dat-pvc hostpath 4m55s

Answer 1

Assumed your sc/pvc/pv are all correct, here's how you can test:
apiVersion: v1
kind: Pod
metadata:
name: dat-tool-pod
labels:
name: dat-tool-pod
spec:
containers:
- name: dat-tool
image: busybox
imagePullPolicy: IfNotPresent
command: ["ash","-c","sleep 7200"]
volumeMounts:
- mountPath: /usr/local/dat/shared
name: dat-volume
restartPolicy: Never
volumes:
- name: dat-volume
persistentVolumeClaim:
claimName: dat-pvc
After the pod is created then you can kubectl exec -it dat-tool-pod -- ash and cd /usr/local/dat/shared. Here you can check the directory/files (incl. permission) to understand why your program complaint about missing directory/files.

Answer 2

For anyone else experiencing this problem, here is what helped me find a solution:
https://github.com/docker/for-win/issues/7023
(And actually the link inside the first comment in this issue.)
So my setup was a Windows 10 machine, using WSL2 to run docker containers and kubernetes cluster on my machine. No matter where I put the folder I wanted to share with my Pod, it didn't appear inside the pod. So based on the link above, I created my folder in /mnt/wsl, called /mnt/wsl/shared.
Because supposedly, this /mnt/wsl folder is where the DockerDesktop will start to look for the folder that you want to share. I changed my PersistentVolume.yml to the following:
apiVersion: v1
kind: PersistentVolume
metadata:
name: dat-volume
spec:
accessModes:
- ReadWriteOnce
capacity:
storage: 150Mi
storageClassName: hostpath
hostPath:
path: /run/desktop/mnt/host/wsl/shared
My understanding is that /run/desktop/mnt/host/wsl is the same as /mnt/wsl, and so I could finally pass files between my Pod and my machine.