I have been trying to run kafka/zookeeper on Kubernetes. Using helm charts I am able to install zookeeper on the cluster. However the ZK pods are stuck in pending state. When I issued describe on one of the pod "didn't find available persistent volumes to bind, 1 node(s) had taints that the pod didn't tolerate.
" was the reason for scheduling failure. But when I issue describe on PVC , I am getting "waiting for first consumer to be created before binding
". I tried to re-spawn the whole cluster but the result is same. Trying to use https://kubernetes.io/blog/2018/04/13/local-persistent-volumes-beta/ as guide.
Can someone please guide me here ?
kubectl get pods -n zoo-keeper
kubectl get pods -n zoo-keeper
NAME READY STATUS RESTARTS AGE
zoo-keeper-zk-0 0/1 Pending 0 20m
zoo-keeper-zk-1 0/1 Pending 0 20m
zoo-keeper-zk-2 0/1 Pending 0 20m
kubectl get sc
kubectl get sc
NAME PROVISIONER AGE
local-storage kubernetes.io/no-provisioner 25m
kubectl describe sc
kubectl describe sc
Name: local-storage
IsDefaultClass: No
Annotations: kubectl.kubernetes.io/last-applied-configuration={"apiVersion":"storage.k8s.io/v1","kind":"StorageClass","metadata":{"annotations":{},"name":"local-storage"},"provisioner":"kubernetes.io/no-provisioner","volumeBindingMode":"WaitForFirstConsumer"}
Provisioner: kubernetes.io/no-provisioner
Parameters: <none>
AllowVolumeExpansion: <unset>
MountOptions: <none>
ReclaimPolicy: Delete
VolumeBindingMode: WaitForFirstConsumer
Events: <none>
kubectl describe pod foob-zookeeper-0 -n zoo-keeper
ubuntu@kmaster:~$ kubectl describe pod foob-zookeeper-0 -n zoo-keeper
Name: foob-zookeeper-0
Namespace: zoo-keeper
Priority: 0
PriorityClassName: <none>
Node: <none>
Labels: app=foob-zookeeper
app.kubernetes.io/instance=data-coord
app.kubernetes.io/managed-by=Tiller
app.kubernetes.io/name=foob-zookeeper
app.kubernetes.io/version=foob-zookeeper-9.1.0-15
controller-revision-hash=foob-zookeeper-5321f8ff5
release=data-coord
statefulset.kubernetes.io/pod-name=foob-zookeeper-0
Annotations: foobar.com/product-name: zoo-keeper ZK
foobar.com/product-revision: ABC
Status: Pending
IP:
Controlled By: StatefulSet/foob-zookeeper
Containers:
foob-zookeeper:
Image: repo.data.foobar.se/latest/zookeeper-3.4.10:1.6.0-15
Ports: 2181/TCP, 2888/TCP, 3888/TCP, 10007/TCP
Host Ports: 0/TCP, 0/TCP, 0/TCP, 0/TCP
Limits:
cpu: 2
memory: 4Gi
Requests:
cpu: 1
memory: 2Gi
Liveness: exec [zkOk.sh] delay=15s timeout=5s period=10s #success=1 #failure=3
Readiness: tcp-socket :2181 delay=15s timeout=5s period=10s #success=1 #failure=3
Environment:
ZK_REPLICAS: 3
ZK_HEAP_SIZE: 1G
ZK_TICK_TIME: 2000
ZK_INIT_LIMIT: 10
ZK_SYNC_LIMIT: 5
ZK_MAX_CLIENT_CNXNS: 60
ZK_SNAP_RETAIN_COUNT: 3
ZK_PURGE_INTERVAL: 1
ZK_LOG_LEVEL: INFO
ZK_CLIENT_PORT: 2181
ZK_SERVER_PORT: 2888
ZK_ELECTION_PORT: 3888
JMXPORT: 10007
Mounts:
/var/lib/zookeeper from datadir (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-nfcfx (ro)
Conditions:
Type Status
PodScheduled False
Volumes:
datadir:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: datadir-foob-zookeeper-0
ReadOnly: false
default-token-nfcfx:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-nfcfx
Optional: false
QoS Class: Burstable
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 69s (x4 over 3m50s) default-scheduler 0/2 nodes are available: 1 node(s) didn't find available persistent volumes to bind, 1 node(s) had taints that the pod didn't tolerate.
kubectl get pv
ubuntu@kmaster:~$ kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
local-pv 50Gi RWO Retain Available local-storage 10m
ubuntu@kmaster:~$
kubectl get pvc local-claim
ubuntu@kmaster:~$ kubectl get pvc local-claim
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
local-claim Pending local-storage 8m9s
ubuntu@kmaster:~$
kubectl describe pvc local-claim
ubuntu@kmaster:~$ kubectl describe pvc local-claim
Name: local-claim
Namespace: default
StorageClass: local-storage
Status: Pending
Volume:
Labels: <none>
Annotations: <none>
Finalizers: [kubernetes.io/pvc-protection]
Capacity:
Access Modes:
VolumeMode: Filesystem
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal WaitForFirstConsumer 2m3s (x26 over 7m51s) persistentvolume-controller waiting for first consumer to be created before binding
Mounted By: <none>
MY PV files:
cat create-pv.yml
apiVersion: v1
kind: PersistentVolume
metadata:
name: local-pv
spec:
capacity:
storage: 50Gi
accessModes:
- ReadWriteOnce
persistentVolumeReclaimPolicy: Retain
storageClassName: local-storage
local:
path: /mnt/kafka-mount
nodeAffinity:
required:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: In
values:
- kmaster
cat pvc.yml
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
name: local-claim
spec:
accessModes:
- ReadWriteOnce
storageClassName: local-storage
resources:
requests:
storage: 50Gi
It looks like you created your PV on master node. By default master node is marked unschedulable by ordinary pods using so called taint. To be able to run some service on master node you have two options:
1) Add toleration to some service to allow it to run on master node:
tolerations:
- effect: NoSchedule
key: node-role.kubernetes.io/master
You may even specify that some service runs only on master node:
nodeSelector:
node-role.kubernetes.io/master: ""
2) You can remove taint from master node, so any pod can run on it. You should know that this is dangerous because can make your cluster very unstable.
kubectl taint nodes --all node-role.kubernetes.io/master-
Read more here and taints and tolerations: https://kubernetes.io/docs/concepts/configuration/taint-and-toleration/