Backup, replication and disaster recovery

Recover the service, not only the files

A resilient recovery design preserves usable points in time, keeps copies beyond the original failure boundary and defines how applications return to operation. Snapshots, replication, backup and disaster recovery each solve a different part of that task.

A recovery path must reach a known usable state
PRODUCTION SERVICE
 
!
 
SNAPSHOTlocal point in time
REPLICAseparate system or site
BACKUPretained recovery history
SELECT → RESTORE → VALIDATE → RETURN THE SERVICE
A snapshot is not automatically a backupIts independence depends on where it is stored and who can alter it.
Replication also copies unwanted changeDeletion or corruption can reach the second system quickly.
Failover does not restore historyA surviving controller can continue serving damaged or encrypted data.
A backup is useful only when it restoresRecovery procedures, credentials and application validation must be tested.
Separate responsibilities

Five methods protect five different objectives

A resilient design combines methods instead of expecting a single mechanism to protect against every failure scenario.

HIGH AVAILABILITY

Continue after a component fails

Redundant paths, controllers or nodes reduce interruption. They do not provide an older copy of the data.

SNAPSHOT

Return to a local point in time

A fast recovery point can help reverse deletion or unwanted change while the source system remains usable.

REPLICATION

Maintain data on another system

Synchronous or scheduled copying protects different recovery objectives and introduces different dependencies.

BACKUP

Retain recoverable history

Backups preserve selected versions according to policy and should remain sufficiently independent from production.

DISASTER RECOVERY

Restore a usable business service

Data, compute, networking, identity, applications and an operational runbook come together at another recovery boundary.

Key distinction: high availability reduces downtime from expected infrastructure failures. Backup and disaster recovery address the ability to return from damaged data, operational mistakes, security incidents or wider outages.

Protection coverage

Choose a recovery control for the event

No single column should be expected to cover the whole table.

Failure or eventHA / failoverSnapshotReplicaIndependent backupDR procedure
Controller, node or path failurePrimary controlNoMay assist recoveryRestore if requiredFor wider interruption
Accidental deletionNoFast local recoveryOnly if an older point remainsHistorical recoveryUsually not required
Application or data corruptionNoIf detected in timeCan copy corruptionSelect a known good pointCoordinates application return
Ransomware or destructive credentialsNoMay share the same risk boundaryMay copy encrypted dataRequires protected, usable copiesContain, recover and validate
Complete system lossOnly if another node survivesNot if stored only on that systemActivate the separate copyRestore to replacement capacityReturns the complete service
Site outageNot within the same siteNot when local onlyIf placed beyond the siteIf a remote copy is availableRequired for dependencies and operations

Coverage depends on the actual placement, retention, permissions and supported product configuration. “Backup”, “replica” or “snapshot” is a design role—not proof that the copy survives a particular event.

Operational recovery

The restore is only one step in the recovery chain

A business service is available again only after the incident is contained and the recovered application is validated.

1

Detect

Confirm the failure, affected services and last known healthy state.

2

Contain

Stop damaged or compromised systems from changing recovery copies.

3

Select

Choose the correct recovery point for the application and its dependencies.

4

Recover

Fail over, activate a replica, restore data or start a recovered virtual machine.

5

Validate

Check data consistency, identity, networking and application behaviour.

6

Return

Reconnect users and monitor the service before closing the incident.

Architecture patterns

Place recovery points across meaningful boundaries

PATTERN 1

Local snapshots for rapid rollback

Fast recovery

Point-in-time recovery on the production storage can minimise transfer time and simplify recovery from recent mistakes.

Production data
Local recovery points
  • Useful for rapid file, dataset or volume recovery
  • Retention must match how long problems may remain undetected
  • A source-system loss can remove both production and local history
PATTERN 2

Scheduled copy to another system or site

Separate boundary

Data is copied at defined intervals to separate capacity. The last successful cycle determines the potential data gap.

Primary storage
Secondary system or location
  • Protects against complete source-system loss
  • Activation, addressing and application dependencies need a runbook
  • Retention is still needed because unwanted change can be copied
PATTERN 3

Independent backup with retention

Historical recovery

Selected recovery points are retained under a separate policy and protected from the same operational or credential failure where possible.

Applications and data
Retained backup history
  • Supports selection of an older known good state
  • Access, deletion protection and credentials are part of the design
  • Regular restore tests confirm that the content is usable
PATTERN 4

Service-level disaster recovery

Business continuity

A documented workflow combines recovery data with compute, networks, identity and application-specific validation.

Unavailable service
Validated alternate service
  • RPO defines the acceptable data gap
  • RTO includes every step until the service is usable
  • Exercises reveal missing dependencies before an incident
Ransomware recovery

Protect recovery options from the same compromise

Ransomware recovery is not simply a larger restore. The team must determine what was affected, prevent reinfection, select a clean point and validate applications before reconnecting users.

  • Keep more than one recovery point so delayed detection does not remove every clean version.
  • Separate administrative access and protect credentials used to manage recovery copies.
  • Place at least one recovery option beyond the production system’s immediate failure boundary.
  • Test complete application recovery—not only whether individual files can be read.

A recoverable copy needs four qualities

Technology alone does not make a copy safe or useful.

Known pointThe team can identify when the data was still usable.
Protected accessThe same compromised identity cannot easily remove every copy.
Independent boundaryOne system or site event does not remove all recovery paths.
Verified restoreApplications, dependencies and data consistency have been exercised.
Virtual machine recovery

Backup & Disaster Recovery is a core part of eEVOS

eEVOS integrates virtualisation with Backup & Disaster Recovery workflows. This keeps protection and recovery close to VM operations while preserving the distinction between keeping a VM available and recovering an earlier usable state.

Instant Backup & Recovery is designed to create VM backups and return protected virtual machines within seconds. It complements the other backup and recovery methods; it does not replace off-system copies or operational validation.

Explore eEVOS backup & DR →

PROTECTED VMBackup policy and retained recovery points
RECOVERED SERVICERestore or Instant Backup & Recovery followed by validation
eEVOS recovery options

Choose the backup path for the recovery objective

eEVOS provides complementary methods for rapid local recovery, separate network and S3-compatible cloud copies, off-site Disaster Recovery and established third-party backup processes.

MethodWhat it protectsDestination or boundaryRecovery value
Instant Backup & RecoveryComplete virtual machines, with file-level recovery when only individual files are requiredeEVOS-managed VM recovery pointsFastest return: backup and whole-VM recovery are designed to complete within seconds
Network BackupEfficient, deduplicated virtual-machine backups, including file-level recoveryA user-selected NAS server reached over the networkSeparate system: keeps a recovery copy beyond the virtualisation host
Cloud Backup (S3-compatible)Space-saving, deduplicated virtual-machine backups, including file-level recoveryAny S3-compatible cloud or object-storage serviceRemote copy: extends protection beyond the local infrastructure
Disaster RecoveryVirtual machines required after a wider infrastructure or site eventAsynchronous replication to a remote location, run on a schedule or manuallyService continuity: VMs can be restored from the off-site location when the primary site is unavailable
External BackupLive virtual-machine images prepared for an organisation’s existing backup processA supported third-party backup tool and its configured repositoryPolicy integration: retains an established backup product where organisational rules require it

Design principle: Instant Backup & Recovery addresses recovery speed. Network, S3-compatible cloud and Disaster Recovery options add different levels of separation from the production environment. The appropriate combination depends on RPO, RTO, retention and failure boundaries.

Implementation examples

euroNAS products protect different layers

The product should be selected by workload, failure boundary and recovery objective—not by the longest feature list.

ProductPrimary roleRelevant protectionBest fit in this designImportant distinction
euroNAS PremiumStandalone storage OS for one server, including Virtual Storage Appliance deploymentSnapshots and scheduled replication to another system or locationRapid storage rollback and a separate recovery copyReplication does not turn one storage server into an automatic local HA pair
euroNAS HA ClusterTwo-node storage with automatic service failoverSynchronous Mirror or shared-storage failover, plus snapshots and scheduled replicationLocal storage availability combined with separate recovery pointsThe synchronous Mirror keeps services available; historical recovery still requires snapshots or another retained copy
eEKASDistributed Ceph scale-out storageReplication or erasure coding across defined failure domainsContinued storage service and data protection during infrastructure failuresDistributed redundancy and self-recovery do not replace a historical backup
eEVOSVirtualisation platformIntegrated Backup & Disaster Recovery, including Instant Backup & Recovery workflowsVM protection, recovery and service return within the virtualisation workflowVM availability and recovery from an earlier state are separate objectives
STORAGE RECOVERY

euroNAS Premium

Snapshots and scheduled replication for standalone storage designs.

Premium details →

AVAILABILITY + RECOVERY

HA Cluster

Automatic local failover with separate snapshot and replication options.

HA Cluster details →

DISTRIBUTED PROTECTION

eEKAS

Ceph protection profiles across nodes and defined failure domains.

eEKAS details →

VM BACKUP & DR

eEVOS

Integrated virtual-machine protection, recovery and Instant Backup & Recovery.

eEVOS backup & DR →

Before selecting the design

Ask operational questions, not only capacity questions

01 · RECOVERY POINTHow much recent data can the business recreate, and how long might a problem remain undetected?
02 · RECOVERY TIMEHow quickly must users receive a validated application—not merely restored storage?
03 · FAILURE BOUNDARYMust recovery survive a drive, node, rack, storage system, administrator account or complete site loss?
04 · APPLICATION CONSISTENCYWhich databases, identity services and multi-VM applications require coordinated recovery?
05 · RETENTIONWhich daily, weekly or longer-term recovery points must remain available for policy and delayed detection?
06 · TESTINGWho performs recovery exercises, how often, and what proves that the service is usable afterwards?
Primary references

Independent guidance for recovery planning

Security guidance and product capabilities change. Validate the complete design and its procedures against current documentation.

SNIA data protection practices

Recovery objectives, data-protection strategy and the operational context around RPO and RTO.

Open SNIA guidance →

CISA #StopRansomware Guide

Organisational practices for preventing, responding to and recovering from ransomware incidents.

Open CISA guidance →

NIST contingency planning

Planning, recovery strategies, testing and maintenance for information-system contingency plans.

Open NIST guidance →

Define recovery objectives before selecting the mechanism

We can review application dependencies, failure boundaries, retention, RPO, RTO and recovery tests for a business storage or virtualisation design.

Discuss your recovery architecture

Scroll to Top