daemonset doesn't create any pods - kubernetes

Question

I'm trying to get going with Kubernetes DaemonSets and not having any luck at all. I've searched for a solution to no avail. I'm hoping someone here can help out.
First, I've seen this ticket. Restarting the controller manager doesn't appear to help. As you can see here, the other kube processes have all been started after the apiserver and the api server has '--runtime-config=extensions/v1beta1=true' set.
kube 31398 1 0 08:54 ? 00:00:37 /usr/bin/kube-apiserver --logtostderr=true --v=0 --etcd_servers=http://dock-admin:2379 --address=0.0.0.0 --allow-privileged=false --portal_net=10.254.0.0/16 --admission_control=NamespaceAutoProvision,LimitRanger,ResourceQuota --runtime-config=extensions/v1beta1=true
kube 12976 1 0 09:49 ? 00:00:28 /usr/bin/kube-controller-manager --logtostderr=true --v=0 --master=http://127.0.0.1:8080 --cloud-provider=
kube 29489 1 0 11:34 ? 00:00:00 /usr/bin/kube-scheduler --logtostderr=true --v=0 --master=http://127.0.0.1:8080
However api-versions only shows version 1:
$ kubectl api-versions
Available Server Api Versions: v1
Kubernetes version is 1.2:
$ kubectl version
Client Version: version.Info{Major:"1", Minor:"2", GitVersion:"v1.2.0", GitCommit:"86327329213fed4af2661c5ae1e92f9956b24f55", GitTreeState:"clean"}
Server Version: version.Info{Major:"1", Minor:"2", GitVersion:"v1.2.0", GitCommit:"86327329213fed4af2661c5ae1e92f9956b24f55", GitTreeState:"clean"}
The DaemonSet has been created, but appears to have no pods scheduled (status.desiredNumberScheduled).
$ kubectl get ds -o json
{
"kind": "List",
"apiVersion": "v1",
"metadata": {},
"items": [
{
"kind": "DaemonSet",
"apiVersion": "extensions/v1beta1",
"metadata": {
"name": "ds-test",
"namespace": "dvlp",
"selfLink": "/apis/extensions/v1beta1/namespaces/dvlp/daemonsets/ds-test",
"uid": "2d948b18-fa7b-11e5-8a55-00163e245587",
"resourceVersion": "2657499",
"generation": 1,
"creationTimestamp": "2016-04-04T15:37:45Z",
"labels": {
"app": "ds-test"
}
},
"spec": {
"selector": {
"app": "ds-test"
},
"template": {
"metadata": {
"creationTimestamp": null,
"labels": {
"app": "ds-test"
}
},
"spec": {
"containers": [
{
"name": "ds-test",
"image": "foo.vt.edu:1102/dbaa-app:v0.10-dvlp",
"ports": [
{
"containerPort": 8080,
"protocol": "TCP"
}
],
"resources": {},
"terminationMessagePath": "/dev/termination-log",
"imagePullPolicy": "IfNotPresent"
}
],
"restartPolicy": "Always",
"terminationGracePeriodSeconds": 30,
"dnsPolicy": "ClusterFirst",
"securityContext": {}
}
}
},
"status": {
"currentNumberScheduled": 0,
"numberMisscheduled": 0,
"desiredNumberScheduled": 0
}
}
]
}
Here is my yaml file to create the DaemonSet
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
name: ds-test
spec:
selector:
app: ds-test
template:
metadata:
labels:
app: ds-test
spec:
containers:
- name: ds-test
image: foo.vt.edu:1102/dbaa-app:v0.10-dvlp
ports:
- containerPort: 8080
Using that file to create the DaemonSet appears to work (I get 'daemonset "ds-test" created'), but no pods are created:
$ kubectl get pods -o json
{
"kind": "List",
"apiVersion": "v1",
"metadata": {},
"items": []
}

Answer 1

(I would have posted this as a comment, if I had enough reputation)
I am confused by your output.
kubectl api-versions should print out extensions/v1beta1 if it is enabled on the server. Since it does not, it looks like extensions/v1beta1 is not enabled.
But kubectl get ds should fail if extensions/v1beta1 is not enabled. So I can not figure out if extensions/v1beta1 is enabled on your server or not.
Can you try GET masterIP/apis and see if extensions is listed there?
You can also go to masterIP/apis/extensions/v1beta1 and see if daemonsets is listed there.
Also, I see kubectl version says 1.2, but then kubectl api-versions should not print out the string Available Server Api Versions (that string was removed in 1.1: https://github.com/kubernetes/kubernetes/pull/15796).

Answer 2

I have this issue in my cluster (k8s version: 1.9.7):
Daemonset controlled by "Daemonset controller" not "Scheduler", So I restart the controller manager, the problem sloved:
But I think this is a issue of kubernetes, some relation info:
Bug 1469037 - Sometime daemonset DESIRED=0 even this matched node
v1.7.4 - Daemonset DESIRED 0 (for node-exporter) #51785

Answer 3

I was facing a similar issue, then tried searching for the daemonset in the kube-system namespace, as mentioned here, https://github.com/kubernetes/kubernetes/issues/61342
I actually did get an output properly as well

Answer 4

For any case that the current state of pods is not equal to desired state (whether it was created by a DaemonSet, ReplicaSet, Deployment etc') I would first check the Kubelet on the current node:
$ sudo systemctl status kubelet
Or:
$ sudo journalctl -u kubelet
In many cases pods weren't created in my cluster because of errors like:
Couldn't parse as pod (Object 'Kind' is missing in 'null')
Which might occur after editing a resource's yaml in editor like vim.

Answer 5

Try:
$kubectl taint nodes --all node-role.kubernetes.io/master:NoSchedule-
master node cannot accept pods.