I want to setup a pre-defined PostgreSQL cluster in a bare meta kubernetes 1.7 with local PV enable. I have three work nodes. I create local PV on each node and deploy the stateful set successfully (with some complex script to setup Postgres replication).
However I'm noticed that there's a kind of naming convention between the volumeClaimTemplates and PersistentVolumeClaim. For example
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
name: postgres
volumeClaimTemplates:
- metadata:
name: pgvolume
The created pvc are pgvolume-postgres-0
, pgvolume-postgres-1
, pgvolume-postgres-2
.
With some tricky, I manually create PVC and bind to the target PV by selector. I test the stateful set again. It seems the stateful set is very happy to use these PVC.
I finish my test successfully but I still have this question. Can I rely on volumeClaimTemplates naming convention? Is this an undocumented feature?
Based on the statefulset API reference
volumeClaimTemplates is a list of claims that pods are allowed to reference. The StatefulSet controller is responsible for mapping network identities to claims in a way that maintains the identity of a pod. Every claim in this list must have at least one matching (by name) volumeMount in one container in the template. A claim in this list takes precedence over any volumes in the template, with the same name.
So I guess you can rely on it.
Moreover, you can define a storage class to leverage dynamic provisioning of persistent volumes, so you won't have to create them manually.
volumeClaimTemplates:
- metadata:
name: www
spec:
accessModes: [ "ReadWriteOnce" ]
storageClassName: my-storage-class
resources:
requests:
storage: 1Gi
Please refer to Dynamic Provisioning and Storage Classes in Kubernetes for more details.