Creating kubernetes statefulset doesnt work

[HannanSolo]

I wanna get Scylla on my GCP kubernetes and am following this guide.

https://github.com/scylladb/scylla-code-samples/tree/master/kubernetes-scylla

but when I run this command kubectl create -f scylla-statefulset.yaml

I get an error

error: error validating "scylla-statefulset.yaml": error validating data: unknown object type schema.GroupVersionKind{Group:"apps", Version:"v1beta2", Kind:"StatefulSet"}; if you choose to ignore these errors, turn validation off with --validate=false

[dorlaor]

@hakuch ^^

[hakuch]

At the time of writing this sample, StatefulSets were in beta.

According to the documentation for StatefulSets, as of release 1.9 they are stable.

If you are running Kubernetes at release 1.9 or greater, I believe you can change apiVersion in scylla-statefulset.yaml to apiVersion: 1.

Does that resolve the error?

[HannanSolo]

@hakuch I am actually trying this on GKE which has a kube version of 1.8.6. So I changed the api version to v1 just for the fun of it and got

error: error validating "scylla-statefulset.yaml": error validating data: unknown object type schema.GroupVersionKind{Group:"", Version:"v1", Kind:"StatefulSet"}; if you choose to ignore these errors, turn validation off with --validate=false

And I'm not sure where to go from there. Also thank you so much for the help, I'm a big fan of your work and the keynotes you did helped a ton.

[hakuch]

@ThePixelBro22, I'm glad to help and appreciate your kind words.

I gave you incorrect advice: try apiVersion: v1.

If that doesn't work, I'll try to spin up a test cluster to try to figure this out definitively.

[HannanSolo]

@hakuch
I get this error

error: error validating "scylla-statefulset.yaml": error validating data: unknown object type schema.GroupVersionKind{Group:"", Version:"v1", Kind:"StatefulSet"}; if you choose to ignore these errors, turn validation off with --validate=false

And this is my yaml.

apiVersion: v1
kind: StatefulSet
metadata:
  name: scylla
  labels:
    app: scylla
spec:
  serviceName: scylla
  replicas: 3
  selector:
    matchLabels:
      app: scylla
  template:
    metadata:
      labels:
        app: scylla
    spec:
      containers:
        - name: scylla
          image: scylladb/scylla:2.0.0
          imagePullPolicy: IfNotPresent
          args: ["--seeds", "scylla-0.scylla.default.svc.cluster.local"]
          ports:
            - containerPort: 7000
              name: intra-node
            - containerPort: 7001
              name: tls-intra-node
            - containerPort: 7199
              name: jmx
            - containerPort: 9042
              name: cql
          resources:
            limits:
              cpu: 500m
              memory: 1Gi
            requests:
              cpu: 500m
              memory: 1Gi
          securityContext:
            capabilities:
              add:
                - IPC_LOCK
          lifecycle:
            preStop:
              exec:
                command: ["/bin/sh", "-c", "PID=$(pidof scylla) && kill $PID && while ps -p $PID > /dev/null; do sleep 1; done"]
          env:
            - name: POD_IP
              valueFrom:
                fieldRef:
                  fieldPath: status.podIP

          readinessProbe:
            exec:
              command:
                - /bin/bash
                - -c
                - exec
                - /opt/ready-probe.sh
            initialDelaySeconds: 15
            timeoutSeconds: 5
          volumeMounts:
            - name: scylla-data
              mountPath: /var/lib/scylla
            - name: scylla-ready-probe
              mountPath: /opt/ready-probe.sh
              subPath: ready-probe.sh
      volumes:
        - name: scylla-ready-probe
          configMap:
            name: scylla
  volumeClaimTemplates:
    - metadata:
        name: scylla-data
        annotations:
          volume.beta.kubernetes.io/storage-class: scylla-ssd
      spec:
        accessModes: ["ReadWriteOnce"]
        resources:
          requests:
            storage: 1Gi

and im using kube 1.8.6

[hakuch]

Based on https://github.com/kubernetes/kubernetes/issues/37393, perhaps you could try apiVersion: apps/v1beta1?

I'm not familiar with Google Kubernetes Engine. Another option is that there is a mismatch between the version of kubectl that you're running locally and the version of Kubernetes that is running on GKE.

[hakuch]

I forgot to mention: I set up a Kubernetes cluster on Google Compute Engine using this guide and specifically installed Kubernetes version 1.8.6 through

$ curl -sS https://get.k8s.io > install.sh
$ chmod +x install.sh
$ KUBERNETES_RELEASE="v1.8.6" ./install.sh

I had no issue creating the StatefulSet using the YAML file as it appears in the repository today.

[HannanSolo]

Ok, ill try it out. Im using google kubernetes engine, on google cloud platform. It just a way more simpler way to creating and managing clusters.

[HannanSolo]

It worked but now i get error on the stateful set on kubernetes.

PersistentVolumeClaim is not bound: "scylla-data-scylla-1" (repeated 3 times)

EDIT

Seems that it works after waiting a bit :)

[hakuch]

Glad to hear it! I think this issue can now be closed.

[HannanSolo]

Argh, so close but yet so far. I'm getting these errors on each pod.

ERROR 2018-01-26 16:59:36,654 [shard 0] init - Bad configuration: invalid value in 'seeds': 'scylla-0.scylla.default.svc.cluster.local': std::system_error (error C-Ares:4, Not found)
ERROR 2018-01-26 16:59:36,654 [shard 0] seastar - Exiting on unhandled exception: bad_configuration_error (std::exception)
INFO  2018-01-26 16:59:36,654 [shard 0] compaction_manager - Asked to stop
INFO  2018-01-26 16:59:36,654 [shard 0] compaction_manager - Stopped
INFO  2018-01-26 16:59:36,655 [shard 1] compaction_manager - Asked to stop
INFO  2018-01-26 16:59:36,655 [shard 1] compaction_manager - Stopped
2018-01-26 16:59:36,669 INFO exited: scylla (exit status 1; not expected)
2018-01-26 16:59:37,671 INFO gave up: scylla entered FATAL state, too many start retries too quickly

Are you getting this as well? @hakuch

EDIT: So I played around a bit and got it to work just by deleting the seed arg, thanks for all the help, Now time to expose the service, any tips?