Glossary/Terminologies

To attach a block device is to make it appear on the Linux node (for example, /dev/longhorn/testvol). If the volume engine dies unexpectedly, SUSE Storage automatically reattaches the volume.

A backup is an object stored in the backupstore. The backupstore may contain both volume backups and system backups.

The backupstore is the external storage location where SUSE Storage backups are saved. It can be either an NFS share, CIFS share, or an S3-compatible object store. SUSE Storage connects to a backupstore through a configured backup target.

A backup target is the endpoint used to access a backupstore in SUSE Storage.

A backup volume represents all backups associated with a single original SUSE Storage volume.

It is stored in the backupstore and visible on the Backup page of the SUSE Storage UI. Backup volumes contain multiple backups for the same volume. Backups are created from snapshots and capture the state of the volume at the time the snapshot was taken. They do not include the snapshot chain or history of changes.

Backups are stored as 2 MiB blocks in object storage by default. For more details on how snapshots and backups work, see the concepts documentation - How Backups Work.

A block-type disk is an unformatted block device that SUSE Storage uses directly for V2 Data Engine volumes.

Cross-cluster disaster recovery allows data from a primary Kubernetes cluster to be recovered on a separate cluster using backups.

A DR volume is a special volume used to maintain a copy of data in a backup cluster so the workload can recover if the primary cluster becomes unavailable.

DR volumes are used to increase the resiliency of SUSE Storage volumes. Each backup volume in the backupstore corresponds to one original volume, and each DR volume corresponds to a backup volume.

DR volumes can be created to accurately reflect backups of a SUSE Storage volume but cannot function as normal volumes until they are activated.

A filesystem-type disk is a block device formatted with an extent-based filesystem (such as ext4 or XFS). SUSE Storage mounts it and stores V1 Data Engine replica data.

Front-end refers to a block device exposed by a SUSE Storage volume.

The SUSE Storage component responsible for managing the lifecycle of engine and replica instances.

A volume attachment mode that attaches the volume without enabling the front-end, primarily used to revert a volume from a snapshot.

A volume expansion performed while the volume is detached.

Overprovisioning allows more logical storage to be allocated than the physical capacity available.

For example, a node with 200 GiB of disk space (with 50 GiB reserved for the operating system) could provision 750 GiB of SUSE Storage volumes. The Storage Over Provisioning Percentage is configurable in the SUSE Storage settings. See Settings for more details.

The replicas of a SUSE Storage volume within the Kubernetes cluster can be considered primary backups.

The read index is an in-memory data structure used by each replica to improve read performance.

It records which differencing disk (snapshot) contains the valid data for each 4K block.

Recurring snapshots allow SUSE Storage to automatically create and retain snapshots at a specified frequency (for example, hourly or daily).

A replica is a SUSE Storage volume data store that consists of a history of the changes in the data within a volume.

A salvage operation is required when all replicas become faulty (for example, due to network disconnection).

During salvage, SUSE Storage tries to identify any usable replicas and then uses them to recover the volume.

Backups stored external to the Kubernetes cluster, on S3, NFS, or CIFS.

A snapshot captures the state of a volume at the time the snapshot is created. Each snapshot stores only the changes that overwrite earlier data, so a chain of snapshots is required to represent the full state. Volumes can be restored from snapshots.

See the concepts documentation - How Snapshots Work.

Snapshot Data Integrity is a SUSE Storage feature that hashes snapshot disk files and periodically checks their integrity to detect filesystem-unaware corruption, such as bit rot.

A system backup contains a bundle of SUSE Storage system resources and is stored in the backupstore.

See the System Backup Bundle documentation for more details on the resources included in a system backup.

SUSE Storage is a thin-provisioned storage system. SUSE Storage volume uses only the space required at a given time.

For example, a 20 GiB volume that stores 1 GiB of data uses only 1 GiB of disk space.

The V1 Data Engine is the original data plane component of SUSE Storage. It uses host filesystem sparse files for thin provisioning, with volumes composed of replicas, supporting snapshots and backups.

The V2 Data Engine is the SUSE Storage data plane. It uses SPDK and huge pages, and block devices to provide higher performance.

The Longhorn Engine is the upstream data plane component corresponding to the SUSE Storage engine.

It is a per-volume storage controller responsible for synchronously replicating data to the volume replicas. This is the underlying term you will see in code, manifests, and CLI outputs.

The Longhorn Manager is the upstream control plane component corresponding to the SUSE Storage manager.

It runs as a Kubernetes DaemonSet and is responsible for managing volumes, handling API requests, and orchestrating Engines, Replicas, and other custom resources. This term is used across internal codebase configurations and manifests.

The default Kubernetes namespace where the control plane, data plane instances, and system-managed components reside.

A Longhorn Volume represents the underlying Kubernetes custom resource (CRD) for a SUSE Storage volume.

For each volume, the Manager creates a Longhorn Engine instance and multiple replicas, each containing a snapshot chain representing the volume history.

The upstream command-line tool used to interact with the system, check data integrity, and manage snapshots. You will see this referenced when running CLI commands.

A storage approach in which data is stored in fixed-size blocks, each identified by a memory address.

A Kubernetes custom resource definition.

The CSI Driver is a container storage interface that can be used with Kubernetes. The CSI driver for SUSE Storage volumes is named driver.longhorn.io.

A Linux filesystem supported by SUSE Storage for storage.

A Linux command used to attach a block device to a directory on the node (for example, mount /dev/longhorn/testvol/mnt).

A distributed filesystem protocol that provides network-based file access.

SUSE Storage supports using NFS in the following cases: * As a backupstore for secondary storage. * As the underlying filesystem for RWX volumes.

A data storage architecture that manages data as objects, each containing the data, a variable amount of metadata, and a globally unique identifier.

SUSE Storage supports backing up volumes to S3-compatible object stores.

A PersistentVolume (PV) is a Kubernetes resource that represents a piece of storage in the cluster.

It may be provisioned manually or dynamically using StorageClasses. It is a cluster-level resource and is required for pods to use persistent storage that is independent of the pod lifecycle.

For more information, see the Kubernetes documentation on persistent volumes.

A PersistentVolumeClaim (PVC) is a user request for storage. Claims specify the desired size and access modes (for example, ReadWriteOnce or ReadOnlyMany).

Pods use PVCs to obtain persistent storage. See the official Kubernetes documentation at persistent volumes.

After reattachment, SUSE Storage automatically detects and mounts the filesystem of the volume.

Amazon S3 is an object storage service.

A network file-sharing protocol that provides remote file access similar to local storage. SUSE Storage supports using SMB/CIFS as a backupstore for secondary storage.

StatefulSets have a stable identity, meaning Kubernetes does not force-delete the Pod.

A Kubernetes resource used for managing stateful applications.

A Kubernetes resource used to automatically provision PersistentVolumes for pods.

A Linux command that detaches a filesystem from the file hierarchy.

A Kubernetes volume allows a pod to store files during its lifetime.

These files persist across container restarts but not when the pod is deleted. To preserve storage beyond the pod lifecycle, a PersistentVolume is required.

See the Kubernetes documentation at Volumes for more details.

A filesystem supported by most Linux distributions. SUSE Storage supports XFS for storage.