In this meetup, Oleg, CTO at Kublr, walks you through the basics of K8s persistence management functionality and how it can be used to simplify managing persistent applications across different environments - in the cloud or on premise. Oleg will use a demo environment with clusters in different clouds to show K8s persistence in practice.
We will cover:
• Persistent data abstractions in K8s: persistent volumes (PV) and their attributes
• PV specifics in different clouds
• Using PV in K8s: persistent volume claims (PVC) and storage classes (SC)
• Automatic volume provisioning
• Persistence and scheduling interrelationships
• Practical examples
Kubernetes (K8s) is a powerful and flexible open source container orchestration system. The power of K8s comes from its modularity and simplicity of basic concepts. Each of these basic concepts build on the other and, from the most basic elements to more advanced ones, each is responsible for its own well-defined logic and behavior.
9. Storage Lifecycle: Provisioning and Use
@olgch; @kublr
Lifecycle Process
Pod Volume 1. Storage is provisioned admin
2. Pod is started with Volume referring to the Storage
Static
Provisioning
1. Storage is provisioned admin
2. PV is created referring to the Storage admin
3. PVC is created referring to the PV user
4. PVC is “bound” to the PV K8S controller
5. Pod is started with Volume referring to the PVC
Dynamic
Provisioning
1. Storage Class is created with parameters for provisioner admin
2. PVC is created referring to a Storage Class object user or K8S
3. Storage is provisioned and PV is created provisioner
4. PVC is “bound” to the PV K8S controller
5. Pod is started with Volume referring to the PVC
10. Storage Lifecycle: Pod Volume
@olgch; @kublr
Kubernetes Cluster
Namespace
Pod
Container 1 Container 2
Volume
Volume
Mount
Volume
Device
“actual”
storage
Spec
1 Provision storage
2
Pod reference storage