Volume Snapshots
Explore this Page
- Overview
- Requirements
- Volume Snapshots for Replicated PV Mayastor
- Volume Snapshots for Local PV ZFS
- Benefits of Volume Snapshots
Overview
Volume Snapshots enable point-in-time, consistent copies of persistent volumes managed by storages such as Replicated PV Mayastor and Local PV ZFS. These snapshots are critical for data protection, backup and restore, disaster recovery, and migration use cases in Kubernetes environments.
DataCore Puls8 supports copy-on-write (COW) based snapshots that capture only the changed blocks, thereby improving storage efficiency. Snapshots can be created, listed, and deleted using standard Kubernetes commands.
Typical use cases include:
- Backup and disaster recovery
- Application cloning and testing
- Rollbacks or troubleshooting
Snapshots are:
- Consistent: Data remains consistent across all volume replicas.
- Immutable: Once created, snapshot data cannot be altered.
Snapshots are not reconstructable in the event of node failure, unlike volume replicas.
Requirements
Before using Volume Snapshots, ensure the following requirements are met:
- Kubernetes cluster with CSI snapshot CRDs installed.
- CSI Snapshot Controller is deployed and running.
- DataCore Puls8 is installed and configured with:
- Replicated PV Mayastor
- Local PV ZFS
- DiskPools are created and operational (For Replicated PV Mayastor).
- A PVC is deployed using a compatible StorageClass.
Volume Snapshots for Replicated PV Mayastor
This section explains how to create, manage, and validate volume snapshots for persistent volumes provisioned by Replicated PV Mayastor. These snapshots are block-level, application-consistent, and support advanced storage capabilities such as incremental backup and thin provisioning.
Creating a StorageClass
cat <<EOF | kubectl create -f -
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: mayastor-3
parameters:
protocol: nvmf
repl: "3"
provisioner: io.openebs.csi-mayastor
EOF
Create and Verify a PVC
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
ms-volume-claim Bound pvc-fe1a5a16-ef70-4775-9eac-2f9c67b3cd5b 1Gi RWO mayastor-3 15s
Copy the PVC name (ms-volume-claim in this case) for later use.
Ensure an application is deployed using the created PVC, as described in the Deploy an Application documentation.
Creating a Volume Snapshot
Snapshots can be created either with an active application or directly from the PVC.
-
Define the VolumeSnapshotClass.
CopyCreate a Default VolumeSnapshotClass for Replicated PV Mayastorcat <<EOF | kubectl create -f -
kind: VolumeSnapshotClass
apiVersion: snapshot.storage.k8s.io/v1
metadata:
name: csi-mayastor-snapshotclass
annotations:
snapshot.storage.kubernetes.io/is-default-class: "true"
driver: io.openebs.csi-mayastor
deletionPolicy: Delete
EOFAlternatively, apply using a YAML file:
-
Create the Snapshot.
CopyCreate a VolumeSnapshot from an Existing PVCcat <<EOF | kubectl create -f -
apiVersion: snapshot.storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
name: mayastor-pvc-snap
spec:
volumeSnapshotClassName: csi-mayastor-snapshotclass
source:
persistentVolumeClaimName: ms-volume-claim
EOFSnapshots on thick-provisioned volumes are automatically converted to thin-provisioned volumes.
Listing Snapshots
Use the following commands to view snapshot and snapshot content details.
NAME READYTOUSE SOURCEPVC SOURCESNAPSHOTCONTENT RESTORESIZE SNAPSHOTCLASS SNAPSHOTCONTENT CREATIONTIME AGE
mayastor-pvc-snap true ms-volume-claim 1Gi csi-mayastor-snapshotclass snapcontent-174d9cd9-dfb2-4e53-9b56-0f3f783518df 57s 57s
NAME READYTOUSE RESTORESIZE DELETIONPOLICY DRIVER VOLUMESNAPSHOTCLASS VOLUMESNAPSHOT VOLUMESNAPSHOTNAMESPACE AGE
snapcontent-174d9cd9-dfb2-4e53-9b56-0f3f783518df true 1073741824 Delete io.openebs.csi-mayastor csi-mayastor-snapshotclass mayastor-pvc-snap default 87s
Deleting a Snapshot
Use the following commands to delete a snapshot.
Filesystem Consistent Snapshots
DataCore Puls8 supports filesystem-consistent snapshots by default. This ensures data integrity by quiescing active I/O operations using the FIFREEZE and FITHAW ioctls during snapshot creation. If the filesystem quiescing process fails, the entire snapshot operation is retried automatically by the Mayastor CSI controller.
To disable filesystem quiescing, modify the VolumeSnapshotClass as follows:
kind: VolumeSnapshotClass
apiVersion: snapshot.storage.k8s.io/v1
metadata:
name: csi-mayastor-snapshotclass
annotations:
snapshot.storage.kubernetes.io/is-default-class: "true"
parameters:
quiesceFs: none
driver: io.openebs.csi-mayastor
deletionPolicy: Delete
Snapshot Capacity and Commitment Considerations
Replicated PV Mayastor enforces capacity admission controls using commitment thresholds to ensure safe and reliable operation of storage pools. These commitment limits apply broadly to volume creation, replica placement, and snapshot operations, and help prevent pool over-allocation and potential data integrity risks.
Understanding these commitment controls is important for capacity planning and for avoiding failures during volume provisioning, replica creation, or snapshot-based backup operations.
Commitment Controls
Replicated PV Mayastor uses the following commitment thresholds to regulate storage allocation:
Pool Commitment
Pool commitment defines the maximum allowed logical overcommitment of a DiskPool when thin provisioning is enabled.
This allows logical allocation to exceed physical capacity within safe limits. For example, if a pool has 10 GiB physical capacity and the pool commitment is configured as 250%, the pool can support up to 25 GiB of logically provisioned storage, subject to other commitment constraints.
If the pool reaches its configured commitment limit, further operations may be blocked, including:
- Volume creation
- Replica creation
- Snapshot creation
This behavior prevents unsafe over-allocation of storage resources.
Volume Commitment
Volume commitment defines the minimum required free space on each replica pool when creating replicas for an existing volume. This ensures that sufficient space is available to safely accommodate replica data and future write operations.
If the required free space is not available on a replica pool, replica creation fails and the volume cannot be provisioned on that pool.
Initial Volume Commitment
Initial volume commitment applies when creating replicas for a new volume. It enforces the same minimum free space requirement during initial volume provisioning to ensure reliable replica placement.
Snapshot Commitment
Snapshot commitment defines the minimum required free space on each replica pool when creating snapshots of an existing volume. This ensures that sufficient space is available to support copy-on-write operations after the snapshot is created.
If any replica pool does not meet the snapshot commitment requirement, snapshot creation fails and any dependent backup operation cannot proceed.
Snapshot Commitment Behavior
VolumeSnapshots require sufficient free space on each replica pool relative to the volume size.
For example:
- Snapshot commitment: 40%
- Volume size: 100 GiB
- Required free space per replica pool: 40 GiB
If any replica pool has less than the required free space, snapshot creation fails.
This check ensures reliable snapshot operation and protects against capacity exhaustion during snapshot lifecycle operations.
Pool Commitment Impact on Snapshots and Thick-Provisioned Volumes
Pool commitment can also affect snapshot creation, particularly for thick-provisioned volumes.
When a snapshot is taken of a thick-provisioned volume:
- The snapshot’s logical size equals the volume size.
- The snapshot initially references existing data using copy-on-write semantics.
- As part of snapshot creation, the volume is internally converted to thin provisioning.
Due to this conversion, the pool must account for potential copy-on-write growth, increasing the pool’s committed capacity by an amount equal to the volume size. If this increase causes the pool to reach its commitment limit, snapshot creation may fail even when sufficient physical free space appears to be available.
This behavior ensures safe storage allocation and prevents capacity exhaustion scenarios.
Example: Snapshot Creation with Commitment Constraints
The following example illustrates how snapshot commitment affects snapshot creation when replica pool capacity is constrained. In this example, volumes are thick-provisioned and snapshot commitment is configured to 40%.
| Volume Size | Free Space per Pool | Required Free Space (40%) | Snapshot Result |
| 7 GiB | 3 GiB | 2.8 GiB | Successful |
| 8 GiB | 2 GiB | 3.2 GiB | Failed |
| 9 GiB | 1 GiB | 3.6 GiB | Failed |
Snapshot creation succeeds only when all replica pools meet the snapshot commitment requirement. If any replica pool fails the check, the snapshot and any dependent backup operation fails.
Default Commitment Values and Configuration
Replicated PV Mayastor uses configurable Helm parameters to enforce commitment controls:
- Pool Commitment
--set openebs.mayastor.agents.core.capacity.thin.poolCommitment- Default: 250%
- Defines the maximum logical overcommitment allowed per DiskPool.
- Volume Commitment
--set openebs.mayastor.agents.core.capacity.thin.volumeCommitment- Default: 40%
- Defines the minimum free space required to create replicas for existing volumes.
- Initial Volume Commitment
--set openebs.mayastor.agents.core.capacity.thin.volumeCommitmentInitial- Default: 40%
- Defines the minimum free space required when provisioning new volumes.
- Snapshot Commitment
--set openebs.mayastor.agents.core.capacity.thin.snapshotCommitment- Default: 40%
- Defines the minimum free space required on each replica pool to allow snapshot creation.
The default values are suitable for most environments and provide a balanced trade-off between capacity utilization and operational safety. In typical deployments, these values do not require modification.
However, environments with large volumes, frequent snapshots, or aggressive thin provisioning may require tuning these parameters during installation or upgrade. Any changes should be accompanied by careful capacity planning and continuous monitoring of DiskPool utilization to ensure reliable snapshot creation.
Volume Snapshots for Local PV ZFS
This section explains the process for creating and managing volume snapshots for Local PV ZFS volumes. ZFS snapshots are file-system level, point-in-time copies that support clone and backup workflows in Kubernetes environments.
Creating a Volume Snapshot
-
Define the VolumeSnapshotClass.
CopyCreate a Default VolumeSnapshotClass for Local PV ZFS$ cat snapshotclass.yaml
kind: VolumeSnapshotClass
apiVersion: snapshot.storage.k8s.io/v1
metadata:
name: zfspv-snapclass
annotations:
snapshot.storage.kubernetes.io/is-default-class: "true"
driver: zfs.csi.openebs.io
deletionPolicy: DeleteAlternatively, apply using a YAML file:
-
Identify the PVC.
CopySample OutputNAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
csi-zfspv Bound pvc-73402f6e-d054-4ec2-95a4-eb8452724afb 4Gi RWO openebs-zfspv 2m35s -
Create the Snapshot.
CopyCreate a VolumeSnapshot$ cat snapshot.yaml
apiVersion: snapshot.storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
name: zfspv-snap
spec:
volumeSnapshotClassName: zfspv-snapclass
source:
persistentVolumeClaimName: csi-zfspvCreate the snapshot in the same namespace as the PVC. Ensure
readyToUse: truebefore using the snapshot.
Listing Snapshots
Use the following commands to view snapshot and snapshot content details.
NAME READYTOUSE SOURCEPVC SOURCESNAPSHOTCONTENT RESTORESIZE SNAPSHOTCLASS SNAPSHOTCONTENT CREATIONTIME AGE
zfspv-snap true csi-zfspv - 4Gi zfspv-snapclass snapcontent-b747cc44-6845-4e72-b0a9-4fb65858e013 106s 106s
apiVersion: snapshot.storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
annotations:
kubectl.kubernetes.io/last-applied-configuration: |
{"apiVersion":"snapshot.storage.k8s.io/v1","kind":
"VolumeSnapshot","metadata":{"annotations":{},"name":
"zfspv-snap","namespace":"default"},"spec":{"source":{"persistentVolumeClaimName":"csi-zfspv"},"volumeSnapshotClassName":"zfspv-snapclass"}}
creationTimestamp: "2020-02-25T08:25:51Z"
finalizers:
- snapshot.storage.kubernetes.io/volumesnapshot-as-source-protection
- snapshot.storage.kubernetes.io/volumesnapshot-bound-protection
generation: 1
name: zfspv-snap
namespace: puls8
resourceVersion: "447494"
selfLink: /apis/snapshot.storage.k8s.io/v1/namespaces/default/volumesnapshots/zfspv-snap
uid: 3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd
spec:
source:
persistentVolumeClaimName: csi-zfspv
volumeSnapshotClassName: zfspv-snapclass
status:
boundVolumeSnapshotContentName: snapcontent-3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd
creationTime: "2020-02-25T08:25:51Z"
readyToUse: true
restoreSize: "0"
Verifying Local PV ZFS Snapshot CR
Check the custom resource created by the ZFS driver.
kubectl get zfssnap snapshot-3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd -n puls8 -o yaml
apiVersion: openebs.io/v1alpha1
kind: ZFSSnapshot
metadata:
creationTimestamp: "2020-02-25T08:25:51Z"
finalizers:
- zfs.openebs.io/finalizer
generation: 2
labels:
kubernetes.io/nodename: e2e1-node2
openebs.io/persistent-volume: pvc-73402f6e-d054-4ec2-95a4-eb8452724afb
name: snapshot-3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd
namespace: puls8
resourceVersion: "447328"
selfLink: /apis/openebs.io/v1alpha1/namespaces/openebs/zfssnapshots/snapshot-3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd
uid: 6142492c-3785-498f-aa4a-569ec6c0e2b8
spec:
capacity: "4294967296"
fsType: zfs
ownerNodeID: e2e1-node2
poolName: test-pool
volumeType: DATASET
status:
state: Ready
Node-Level Validation
Validate snapshot presence directly on the node.
NAME USED AVAIL REFER MOUNTPOINT
test-pool 818K 9.63G 24K /test-pool
test-pool/pvc-73402f6e-d054-4ec2-95a4-eb8452724afb 24K 4.00G 24K /var/lib/kubelet/pods/3862895a-8a67-446e-80f7-f3c18881e391/volumes/kubernetes.io~csi/pvc-73402f6e-d054-4ec2-95a4-eb8452724afb/mount
test-pool/pvc-73402f6e-d054-4ec2-95a4-eb8452724afb@snapshot-3cbd5e59-4c6f-4bd6-95ba-7f72c9f12fcd 0B - 24K -
Benefits of Volume Snapshots
- Enables point-in-time backup for critical applications.
- Facilitates fast and reliable disaster recovery.
- Minimizes downtime by enabling quick volume restores.
- Simplifies data migration across environments or clusters.
Learn More