如何在 OSX 上将 Kubernetes Persistent Local Volumes 与 Minikube 一起使用？

Question

I want to experiment with Persistent Local Volumes using Minikube on OSX. I am using the local provisioner ( https://github.com/kubernetes-sigs/sig-storage-local-static-provisioner ) to create the PVs for me.

I create the local directory on the OSX host and mount it to the Minikube VM:

mkdir -p /path/to/data/dir
minikube mount /path/to/data/dir:/data/dir &

When I look at the logs for the provisioner, I can see errors like this:

Found new volume at host path "/data/dir" with capacity 0, creating Local PV "local-pv-ccc90d9b", required volumeMode "Filesystem"
Error creating PV "local-pv-ccc90d9b" for volume at "/data/zk-2": PersistentVolume "local-pv-ccc90d9b" is invalid: spec.capacity[storage]: Invalid value: "0": must be greater than zero

Guided by the advice here ( https://github.com/kubernetes-incubator/external-storage/issues/968 ), I ssh'ed in to the Minikube VM and looked at the capacity of the mount:

$ df /data/dir/
Filesystem     1K-blocks  Used Available Use% Mounted on
192.168.99.1           0     0         0    - /data/dir

So the capacity of the mount is zero and this seems to the why it's throwing up the error I see. However, the parent directory

$ df /data          
Filesystem     1K-blocks    Used Available Use% Mounted on
/dev/sda1       16888216 1129408  14769676   8% /data

Is this an issue with the minikube mount command? Can this be made to work on OSX? How do I get a Persistent Local Volume to work on OSX Minikube?

Answer 1

Here is my little how-to for those who ended up on that page looking for a way to share local folders on their Mac with minikube.

Personally, I was spinning up some Azure SQL Edge server (an equivalent to MS SQL Server, but that is compiled for the arm architecture of M1 silicon chips) and I wanted the database's data to persist on my disk so I could restart my pods without loosing their data.

1. Mount the folder you want to share on your host, in minikube:

minikube mount ./path/to/mySharedData:/mnt1/shared1 --uid 10001 --gid 10001

Here, the volume mounted in minikube will have group id and user id 10001. This is the user id of Azure SQL Edge server inside the container.

I used these --uid and --gid arguments because I was running into some write access error.

If you don't know what is the user id inside your container, log into your container and type id , it will tell you the user id.

Don't close the terminal. That process needs to be running all the time for the folder to be accessible.

2. Use that folder with hostPath:

    spec:
      volumes:
        - name: mssql-data
          hostPath:
            path: /mnt1/shared1
      containers:
        - name: azuresqledge
          image: mcr.microsoft.com/azure-sql-edge:latest
          ports:
            - containerPort: 1433
          volumeMounts:
            - name: mssql-data
              mountPath: /var/opt/mssql/data
          env:
            - name: MSSQL_PID
              value: "Developer"
            - name: ACCEPT_EULA
              value: "Y"
            - name: SA_PASSWORD
              valueFrom:
                secretKeyRef:
                  name: mssql
                  key: SA_PASSWORD
            - name: MSSQL_AGENT_ENABLED
              value: "TRUE"
            - name: MSSQL_COLLATION
              value: "SQL_Latin1_General_CP1_CI_AS"
            - name: MSSQL_LCID
              value: "1033"
      terminationGracePeriodSeconds: 30
      securityContext:
        fsGroup: 10001

A lot of the stuff here is Azure SQL specific (all the env ...). Two things are important: Define the volume mounted in minikube with:

      volumes:
        - name: mssql-data
          hostPath:
            path: /mnt1/shared1

And mount it inside your container where you need:

          volumeMounts:
            - name: mssql-data
              mountPath: /var/opt/mssql/data