Volume issue when installing Prometheus using Helm chart

General Discussions
1

I have a Kubernetes cluster comprised of 1 control plane node and 2 worker nodes. The cluster was bootstrapped using kubeadm.

I am trying to install Prometheus in this cluster using Helm charts as follows:

helm install prometheus prometheus-community/prometheus

However, I am running into the following problem: prometheus-server and prometheus-alertmanager pods remain pending.

I followed this tutorial to explictly add volume definitions as follows:

kubectl create -f - <<EOF
kind: PersistentVolume
apiVersion: v1
metadata:
  name: pv-prom-server
spec:
  storageClassName: 
  capacity:
    storage: 5Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/mnt/pv_prom_server"
---
kind: PersistentVolume
apiVersion: v1
metadata:
  name: pv-prom-alertmgr
spec:
  storageClassName: 
  capacity:
    storage: 5Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/mnt/pv_prom_altermgr"
EOF

But I end up with the following when I perform kubectl get pv. pv-prom-server seems to be bound to prometheus-alertmanager!

NAME               CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS      CLAIM                             STORAGECLASS   REASON   AGE
pv-prom-alertmgr   5Gi        RWO            Retain           Available                                                             6m11s
pv-prom-server     5Gi        RWO            Retain           Bound       default/prometheus-alertmanager                           6m11s

prometheus-alertmanager turns into running but prometheus-server remains pending.

Is there a way to solve this issue?

Thanks!

2

I have the same issue. Please help !

3

Hi:) Could you please share how did you resolve this issue? I have looked at the screenshot and there seems to be no problem in prom-server being bound to alert manager, that is the goal I believe but correct me if I am mistaken.