Storage classes for dynamic provisioning in Kubernetes - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a StorageClass in Kubernetes?

A StorageClass defines how a unit of storage is dynamically provisioned in Kubernetes. It specifies the type of storage (like SSD or HDD), the provisioner, and parameters for creating persistent volumes automatically.

beginner

How does dynamic provisioning work with StorageClasses?

When a PersistentVolumeClaim requests storage with a StorageClass, Kubernetes uses the StorageClass's provisioner to create a PersistentVolume automatically without manual admin intervention.

intermediate

What key fields are defined in a StorageClass YAML?

Key fields include:
- provisioner: the plugin that creates storage
- parameters: settings for the storage
- reclaimPolicy: what happens to storage after release
- volumeBindingMode: when volume binding happens

intermediate

What is the purpose of the reclaimPolicy in a StorageClass?

The reclaimPolicy tells Kubernetes what to do with the storage after the PersistentVolumeClaim is deleted. Common values are Delete (storage is deleted) or Retain (storage is kept for manual cleanup).

beginner

Give an example of a common provisioner used in StorageClasses.

A common provisioner is kubernetes.io/aws-ebs for AWS Elastic Block Store, or kubernetes.io/gce-pd for Google Cloud Persistent Disk. These provisioners create cloud storage volumes dynamically.

What does a StorageClass in Kubernetes define?

A Network policies for pods

B User permissions for storage access

C How storage volumes are dynamically created

D Container image versions

A StorageClass defines how storage volumes are dynamically provisioned in Kubernetes.

Which field in a StorageClass specifies the plugin to create storage?

A parameters

B reclaimPolicy

C volumeBindingMode

D provisioner

The provisioner field specifies the plugin that creates the storage volume.

What happens if reclaimPolicy is set to Delete?

A Storage volume is kept after claim deletion

B Storage volume is deleted after claim deletion

C Storage volume is backed up automatically

D Storage volume is resized automatically

Delete means the storage volume is deleted when the PersistentVolumeClaim is deleted.

Dynamic provisioning eliminates the need for:

A Manual creation of PersistentVolumes

B Using PersistentVolumeClaims

C Defining StorageClasses

D Pod scheduling

Dynamic provisioning automatically creates PersistentVolumes, so admins don't create them manually.

Which of these is NOT a typical parameter in a StorageClass?

A pod CPU limits

B type of disk (SSD/HDD)

C provisioner plugin name

D reclaim policy

Pod CPU limits are unrelated to StorageClass parameters.

Explain how StorageClasses enable dynamic provisioning in Kubernetes.

Think about how storage is created automatically when requested.

Describe the role of reclaimPolicy in a StorageClass and its common values.

Consider what happens to storage after you stop using it.

Practice

1. What is the main purpose of a StorageClass in Kubernetes?

easy

A. To define how storage volumes are dynamically provisioned

B. To schedule pods on specific nodes

C. To manage network policies

D. To configure container resource limits

Solution

1. Step 1: Understand StorageClass role

   A StorageClass defines parameters for dynamic volume provisioning in Kubernetes.

2. Step 2: Compare with other options

   Scheduling pods, managing network policies, and resource limits are unrelated to StorageClass.

3. Final Answer:

   To define how storage volumes are dynamically provisioned -> Option A

4. Quick Check:

   StorageClass = dynamic provisioning [OK]

Hint: StorageClass controls storage creation, not pods or network [OK]

Common Mistakes:

• Confusing StorageClass with pod scheduling
• Thinking StorageClass manages network policies
• Assuming StorageClass sets container CPU/memory limits

2. Which of the following is the correct YAML snippet to specify a StorageClass named fast-storage for a PersistentVolumeClaim?

easy

A. StorageClass: fast-storage

B. storageclass: fast-storage

C. storageClassName: fast-storage

D. storage_class_name: fast-storage

Solution

1. Step 1: Recall correct field name

   The correct field in PVC spec is storageClassName with camelCase.

2. Step 2: Check other options

   Options B, C, and D use incorrect casing or underscores, which are invalid in Kubernetes YAML.

3. Final Answer:

   storageClassName: fast-storage -> Option C

4. Quick Check:

   Correct YAML key = storageClassName [OK]

Hint: Use exact camelCase 'storageClassName' in PVC spec [OK]

Common Mistakes:

• Using wrong casing like 'storageclass' or 'StorageClass'
• Replacing camelCase with underscores
• Misspelling the field name

3. Given this StorageClass YAML snippet:

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: slow
provisioner: kubernetes.io/no-provisioner
volumeBindingMode: WaitForFirstConsumer
reclaimPolicy: Retain

What will happen when a PersistentVolumeClaim requests storage from this class?

medium

A. The volume will not be dynamically provisioned; manual provisioning is required

B. The volume will be deleted automatically when released

C. The volume will be dynamically provisioned immediately

D. The volume will be provisioned but bound immediately without waiting

Solution

1. Step 1: Analyze provisioner field

   The provisioner is kubernetes.io/no-provisioner, meaning no dynamic provisioning.

2. Step 2: Understand reclaimPolicy and binding mode

   ReclaimPolicy is Retain (volume kept after release), and volumeBindingMode waits for pod scheduling.

3. Final Answer:

   The volume will not be dynamically provisioned; manual provisioning is required -> Option A

4. Quick Check:

   No provisioner = manual provisioning [OK]

Hint: No provisioner means no dynamic provisioning [OK]

Common Mistakes:

• Assuming dynamic provisioning happens with no-provisioner
• Confusing reclaimPolicy with provisioning behavior
• Ignoring volumeBindingMode effect

4. You created a StorageClass with volumeBindingMode: Immediate but your pods fail to schedule because volumes are not available in the node's zone. What is the likely cause and fix?

medium

A. Cause: StorageClass name mismatch; Fix: Correct PVC storageClassName

B. Cause: Reclaim policy is Retain; Fix: Change to Delete

C. Cause: Provisioner is missing; Fix: Add a provisioner

D. Cause: Immediate binding ignores pod location; Fix: Change to WaitForFirstConsumer

Solution

1. Step 1: Understand volumeBindingMode impact

   Immediate binding provisions and binds volume before pod scheduling, ignoring node zone.

2. Step 2: Identify fix

   Changing to WaitForFirstConsumer delays binding until pod scheduling, ensuring volume is in correct zone.

3. Final Answer:

   Cause: Immediate binding ignores pod location; Fix: Change to WaitForFirstConsumer -> Option D

4. Quick Check:

   Immediate binding ignores pod zone [OK]

Hint: Use WaitForFirstConsumer to bind volume after pod scheduling [OK]

Common Mistakes:

• Changing reclaimPolicy instead of binding mode
• Adding provisioner when already present
• Fixing storageClassName without checking binding mode

5. You want to create a StorageClass that dynamically provisions SSD-backed volumes with fast performance, deletes volumes when released, and delays binding until pod scheduling. Which YAML snippet correctly implements this?

hard

A. apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast-ssd
provisioner: kubernetes.io/aws-ebs
parameters:
type: standard
reclaimPolicy: Retain
volumeBindingMode: Immediate

B. apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast-ssd
provisioner: kubernetes.io/aws-ebs
parameters:
type: gp2
reclaimPolicy: Delete
volumeBindingMode: WaitForFirstConsumer

C. apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast-ssd
provisioner: kubernetes.io/no-provisioner
parameters:
type: io1
reclaimPolicy: Delete
volumeBindingMode: WaitForFirstConsumer

D. apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: fast-ssd
provisioner: kubernetes.io/aws-ebs
parameters:
type: io1
reclaimPolicy: Delete
volumeBindingMode: Immediate

Solution

1. Step 1: Identify correct provisioner and parameters

   For AWS SSD, provisioner is kubernetes.io/aws-ebs and type gp2 is general SSD.

2. Step 2: Check reclaimPolicy and binding mode

   ReclaimPolicy should be Delete to remove volumes after release; binding mode WaitForFirstConsumer delays binding until pod scheduling.

3. Step 3: Validate options

   apiVersion: storage.k8s.io/v1 kind: StorageClass metadata: name: fast-ssd provisioner: kubernetes.io/aws-ebs parameters: type: standard reclaimPolicy: Retain volumeBindingMode: Immediate uses wrong type and policies; apiVersion: storage.k8s.io/v1 kind: StorageClass metadata: name: fast-ssd provisioner: kubernetes.io/no-provisioner parameters: type: io1 reclaimPolicy: Delete volumeBindingMode: WaitForFirstConsumer uses no provisioner; apiVersion: storage.k8s.io/v1 kind: StorageClass metadata: name: fast-ssd provisioner: kubernetes.io/aws-ebs parameters: type: io1 reclaimPolicy: Delete volumeBindingMode: Immediate uses Immediate binding which is not desired; apiVersion: storage.k8s.io/v1 kind: StorageClass metadata: name: fast-ssd provisioner: kubernetes.io/aws-ebs parameters: type: gp2 reclaimPolicy: Delete volumeBindingMode: WaitForFirstConsumer matches all requirements.

4. Final Answer:

   YAML with provisioner kubernetes.io/aws-ebs, type gp2, reclaimPolicy Delete, and WaitForFirstConsumer -> Option B

5. Quick Check:

   Correct provisioner + Delete + WaitForFirstConsumer = apiVersion: storage.k8s.io/v1 kind: StorageClass metadata: name: fast-ssd provisioner: kubernetes.io/aws-ebs parameters: type: gp2 reclaimPolicy: Delete volumeBindingMode: WaitForFirstConsumer [OK]

Hint: Match provisioner, parameters, reclaimPolicy, and bindingMode exactly [OK]

Common Mistakes:

• Using no-provisioner disables dynamic provisioning
• Choosing wrong volume type parameter
• Setting reclaimPolicy to Retain instead of Delete
• Using Immediate binding when delayed binding is needed