Cannot start a stopped kind cluster

[awk]

If I create a cluster with kind:

vctl system start
Preparing storage...
Container storage has been prepared successfully under /Users/awk/.vctl/storage
Launching container runtime...
Container runtime has been started.

vctl kind
vctl-based KIND is ready now. KIND will run local Kubernetes clusters by using vctl containers as "nodes"
* All Docker commands has been aliased to vctl in the current terminal. Docker commands performed in current window would be executed through vctl. If you need to use regular Docker commands, please use a separate terminal window.

bash-3.2$ kind create cluster --name test
Creating cluster "test" ...
 ✓ Ensuring node image (kindest/node:v1.18.2) 🖼 
 ✓ Preparing nodes 📦  
 ✓ Writing configuration 📜 
 ✓ Starting control-plane 🕹️ 
 ✓ Installing CNI 🔌 
 ✓ Installing StorageClass 💾 
Set kubectl context to "kind-test"
You can now use your cluster with:

kubectl cluster-info --context kind-test

Not sure what to do next? 😅  Check out https://kind.sigs.k8s.io/docs/user/quick-start/

bash-3.2$ vctl ps
────                 ─────                                                                                  ───────                   ──                ─────            ──────    ─────────────               
NAME                 IMAGE                                                                                  COMMAND                   IP                PORTS            STATUS    CREATION TIME               
────                 ─────                                                                                  ───────                   ──                ─────            ──────    ─────────────               
test-control-plane   kindest/node@sha256:7b27a6d0f2517ff88ba444025beae41491b016bc6af573ba467b70c5e8e0d85f   /usr/local/bin/entry...   192.168.244.132   62541:6443/tcp   running   2020-11-14T12:08:38-05:00   
bash-3.2$ kubectl cluster-info --context kind-test
Kubernetes master is running at https://127.0.0.1:62541
KubeDNS is running at https://127.0.0.1:62541/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy

To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.

Everything looks great, now if I stop the node/container:

bash-3.2$ vctl stop test-control-plane
WARNING unknown stop signal: SIGRTMIN+3, using default 
WARNING container test-control-plane did not stop within 10 second, force kill it... 
INFO container test-control-plane has been stopped 
bash-3.2$ vctl ps
────   ─────   ───────   ──   ─────   ──────   ─────────────   
NAME   IMAGE   COMMAND   IP   PORTS   STATUS   CREATION TIME   
────   ─────   ───────   ──   ─────   ──────   ─────────────   

And then attempt to restart it - it fails:

bash-3.2$ vctl start test-control-plane
ERROR error checking volume mapping config: mkdir /lib: read-only file system 

Is it not intended that clusters can be restarted? That they must be thrown away and new ones created if there's a power failure or some other system restart?

[jeffwubj]

Thanks for the reporting, this should be a valid issue. We have fixed it in house and will publish it in a future release and once it has been released, we will mark this issue closed.

[awk]

Great News!

Happy to beta test something if you need someone :-) Check with @mikeroySoft - I'm no stranger to beta builds of Fusion :-)

[jeffwubj]

Great News!

Happy to beta test something if you need someone :-) Check with @mikeroySoft - I'm no stranger to beta builds of Fusion :-)

wow, great to have your help! :-)

[agergec]

Hello Friends,

I have this issue on Fusion 12.1.2.
What is the suggested way to properly stop and start a cluster until the fix is released?

I asked the same question on VMWare communities also.
https://communities.vmware.com/t5/VMware-Fusion-Discussions/How-to-start-the-kind-cluster-after-vctl-system-stop-command/m-p/2858944/highlight/false#M174923

Thank and regards
Gökalp

[jpapejr]

@jeffwubj "We have fixed it in house and will publish it in a future release"

-- can I haz the fix now, plz? It's the future. ;)

[agergec]

@jpapejr Hi, it seems minicube is a good alternative.
It has vmware driver and can join the vmnet8 network. You may give it a try.

[jpapejr]

If I wanted to run minikube I wouldn’t be here messing with kind and vctl, right? ;)

[agergec]

I spent soo much time to find a solution, before running minikube.
I just tried to save your time :)