Question
DOKS - How to Create A Pod in a node where PV (do-block-storage) is mounted
- Posted on January 10, 2023
- DigitalOcean Managed KubernetesKubernetes
Asked by Sisir Kanti Adhikari
DO Block storage only can mount into a single Node (VM). So, when I provision PV/PVC it creates a block storage.
I am implementing helm chart bitnami/mysql where I want to create the mysql pod only into the node the PVC is mounted.
How should I go about it? I have tried with Affinity but I have some issues with it as I can not label nodes. Here is my configuration values.yml file
global:
storageClass: do-block-storage
architecture: standalone
primary:
name: primary
persistence:
enabled: true
existingClaim: backend-mysql-primary
accessModes:
- ReadWriteOnce
size: 10Gi
selector: {}
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- topologyKey: topology.kubernetes.io/region
namespaces:
- sportsdemy
labelSelector:
matchExpressions:
- key: pvc
operator: In
values:
- backend-mysql-primary
This config throws the following error:
error validating "": error validating data: ValidationError(StatefulSet.spec.template.spec.affinity.podAffinity.requiredDuringSchedulingIgnoredDuringExecution[0]): missing required field "topologyKey" in io.k8s.api.core.v1.PodAffinityTerm
So it looks like a topologyKey is required but I am not sure how it is helpful here.
But I went ahead and tried to add the topologyKey from labels assigned to the node by DOKS. topology.kubernetes.io/region
Deployment is applying without issues but pod is assigned to any node:
0/2 nodes are available: 2 node(s) didn't match pod affinity rules. preemption: 0/2 nodes are available: 2 Preemption is not helpful for scheduling.
pod didn't trigger scale-up: 1 node(s) didn't match pod affinity rules
Here is the yml file for volumes:
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: backend-mysql-primary
namespace: sportsdemy
labels:
pvc: backend-mysql-primary
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 10Gi
storageClassName: do-block-storage
Goal: I have labeled my pv as pvc: backend-mysql-primary and want to a node where I have the PV.
This textbox defaults to using Markdown to format your answer.
You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!
These answers are provided by our Community. If you find them useful, show some love by clicking the heart. If you run into issues leave a comment, or add your own answer to help others.
Hollie's Hub for Good
Working on improving health and education, reducing inequality, and spurring economic growth? We’d like to help.
Hi there,
The issue here is that Kubernetes is trying to schedule your Pod based on the affinity rules, but it can’t find any nodes that match the criteria. In your case, you’re trying to schedule the Pod on the node where the Persistent Volume Claim (PVC) is bound, but this is not directly possible with affinity rules because these rules operate based on labels on the nodes or other pods, not based on PVCs.
However, you can achieve something similar by using a workaround:
pvc: backend-mysql-primary. You can find the node where the PV is attached by describing the PV and looking for theNodeAffinityfield.nodeAffinityrule in the Pod spec to schedule the Pod on the node with the label from step 2. Here’s an example of what the affinity part of your values file might look like:Please note that this will only work if the PVC is already bound to the PV and the node has been labeled before you deploy the Helm chart. If you’re dynamically provisioning the PVC along with the Helm chart, you would need to ensure that the PVC is created and bound, and the node is labeled, before the Pod is scheduled.
Also, be aware that if the node goes down, the Pod won’t be able to be scheduled on another node, because the PV can only be attached to the node it was initially bound to.
You might want to consider using a Managed MySQL database cluster for a worry-free setup:
Best,
Bobby