Modern ransomware crews don’t just encrypt your servers. They go looking for your backups first. If an attacker with stolen admin credentials can delete or encrypt your backup repository, you no longer have a recovery plan, you have a ransom negotiation. Immutable backups exist to remove that option: once data is written, nobody, including a compromised administrator, can alter or delete it until a defined retention period ends.
This guide explains how ransomware-proof backups actually work, the difference between the main approaches to immutable backup storage, and how to design a recovery strategy you can trust. If you’re still choosing backup tooling, our overview of 10 backup solutions for dedicated servers is a good starting point. This post focuses on making those backups survive an attack.
Why Ordinary Backups Fail Against Ransomware
- Backups are reachable from the network you’re defending. A backup server joined to the same domain, or a mounted share, can be encrypted or wiped by the same attacker who got into production.
- Credentials are the weak link. If one set of admin credentials can both run production and delete backups, one compromise ends both.
- Attackers deliberately target recovery. Deleting snapshots, shadow copies and backup catalogs is a standard early step in many attacks.
- Dormant infections. Malware can sit unnoticed for weeks. If your retention is shorter than the dwell time, every restore point may already be contaminated.
What “Immutable” Actually Means
An immutable backup is write-once, read-many (WORM): after it’s written, it cannot be modified or deleted before its retention date, regardless of who asks. The guarantee is only as strong as where it’s enforced. Immutability enforced by the storage system itself, below the level any server administrator can override, is meaningfully stronger than a flag a root user can simply remove.
The Main Ways to Get Immutability
1. Object Lock on S3-compatible storage
Object lock storage is the most widely used approach. S3 Object Lock applies a retention period to object versions, and it is supported by many S3-compatible services and self-hosted systems as well. It has two modes that behave very differently:
- Governance mode — protected from ordinary deletion, but users with special permissions can override it. Useful for testing and for environments that need an escape hatch.
- Compliance mode — no one can shorten or remove the retention period, not even the account owner. This is the strongest protection and also the least forgiving: a mistaken retention period or an unwanted large upload stays (and keeps costing you) until it expires.
Object Lock requires versioning to be enabled and should be configured when the bucket is created, so decide your retention approach before the first backup runs.
2. Hardened Linux repositories
Some backup products can write to a hardened Linux repository that applies immutability flags at the filesystem level, with a deliberately locked-down host: no interactive SSH after setup, minimal services, and credentials the production environment never holds. This can be a good fit for on-premises or dedicated-server deployments where you control the hardware.
3. Append-only backup repositories
Tools such as BorgBackup and restic (via its REST server) support append-only modes, where the client can add new backups but cannot delete or rewrite existing ones. The remote side, kept under separate credentials, performs pruning. This is protection by access design rather than storage-level WORM, so the security of the backup host and its admin credentials matters a great deal.
4. Snapshot holds and offline copies
Filesystem snapshots with holds, tape, and removable or offline media provide additional layers. Air-gapped backups, meaning copies with no standing network path to production, protect against attackers who can reach everything online. The trade-off is operational: offline copies are slower to refresh and slower to restore.
Designing the Strategy Around the Technology
Follow 3-2-1-1-0
- 3 copies of your data
- 2 different storage types
- 1 copy offsite
- 1 copy immutable or air-gapped
- 0 errors on tested restores
Set a backup retention policy that outlasts the dwell time
A good backup retention policy balances storage cost against how far back you might need to go. Short daily retention alone is risky if an infection can go unnoticed for weeks. A common approach is tiered retention (for example, recent daily points, weekly points for a few months, monthly points for longer) with the immutable window covering at least your realistic detection time. Compliance requirements may also dictate minimums. Remember that in compliance mode, you can’t shorten retention later.
Separate the credentials and the trust boundary
- Use different accounts, ideally a different identity system, for backup administration than for production.
- Require MFA on every backup console and storage account.
- Give backup agents write-only or append-only permissions where possible.
- Store backup credentials the way you’d store any high-value secret, as covered in our guide to secrets management for production servers.
- Don’t leave backup storage mounted on production hosts.
Encrypt backups and protect the keys
Client-side encryption prevents a storage provider or a stolen repository from exposing your data. Keep encryption keys separate from the backups, and back up the keys themselves. A backup you can’t decrypt is not a backup.
What Immutability Does Not Do
- It doesn’t stop you backing up bad data. If ransomware has already encrypted files before a backup runs, an immutable copy faithfully preserves the encrypted version. You need multiple restore points and ideally anomaly detection on backup sizes and change rates.
- It doesn’t prevent data theft. Many attacks now include exfiltration for extortion. Immutability protects availability, not confidentiality; encryption and access control cover the rest.
- It doesn’t make restores fast. Large datasets restored over limited bandwidth can take days. Plan recovery time, not just recovery points.
- It isn’t a substitute for security hygiene. Patching, hardening and monitoring remain essential, as covered in our list of top tools for enhancing server security.
Test Restores or You Don’t Have Backups
The most common failure in real incidents isn’t missing backups, it’s backups that can’t be restored in time, or at all. Build recovery testing into your routine:
- Restore random files from different dates every month.
- Do a full-system restore into an isolated environment at least quarterly.
- Time it and compare against your recovery time objective (RTO) and recovery point objective (RPO).
- Document the runbook so someone other than the original engineer can perform it under pressure.
- Rehearse the “everything is down and our admin accounts are compromised” scenario, including how you’d get to your backups.
This is the heart of real disaster recovery. For architecture-level resilience that reduces how often you’d need a full restore, see our guide to high availability architecture for business websites. Government guidance such as CISA’s #StopRansomware resources also emphasizes offline, encrypted, tested backups as a core control.
Monitoring and Alerting on the Backup System
- Alert on failed or missing backup jobs, not just on successes.
- Alert on unusual deletion attempts, retention changes and new credentials.
- Watch backup size and change-rate trends, since a sudden spike can be an early sign of mass encryption.
- Forward backup-system audit logs to a separate, protected location.
How BeStarHost Supports Resilient Server Recovery
Backups are one part of a recovery plan, and the infrastructure underneath matters too. BeStarHost’s dedicated server platform provides:
- Dedicated servers with guaranteed, unshared resources, so recovery and restore jobs aren’t competing with other tenants.
- Tier 3 / Tier 4 hardware with RAID 0 / RAID 1 configurations and 99.9% uptime. RAID is protection against disk failure, not a backup.
- IPMI KVM-over-IP for out-of-band access when an operating system needs to be reinstalled or recovered.
- Dedicated, unshared bandwidth on a global low-latency network, useful for moving large restores.
- 14 global data center locations across Europe (France, Germany, Netherlands, United Kingdom), Asia (Singapore, Hong Kong, India, South Korea, Taiwan, Philippines, Myanmar, Cambodia) and North America (United States, Canada), making geographic separation between production and backup copies practical.
- No setup fees and 24/7/365 support.
Explore our dedicated server plans, read more on our About Us page, or contact our team to discuss your recovery requirements.
Frequently Asked Questions
What are immutable backups?
Immutable backups are write-once copies that cannot be modified or deleted until a set retention period expires, even by administrators. This prevents ransomware or a compromised account from destroying your recovery data.
What is the difference between governance mode and compliance mode in Object Lock?
Governance mode protects objects from normal deletion but allows users with special permissions to override the lock. Compliance mode allows no one, including the account owner, to remove or shorten the retention period, giving the strongest protection but little room for error.
Are air-gapped backups the same as immutable backups?
No. Air-gapped backups have no standing network connection to production, while immutable backups can be online but cannot be altered. They protect against different attack paths, and combining both gives the strongest defense.
How long should immutable backups be retained?
Long enough to outlast realistic detection time for an infection, plus any compliance requirements. Many organizations use tiered retention, keeping recent daily points alongside weekly and monthly points, with the immutable window covering at least how long an attack could go unnoticed.
Can immutable backups still be a problem if ransomware encrypted data before the backup ran?
Yes. An immutable backup preserves whatever was written, including already-encrypted files. Keeping multiple older restore points and monitoring backup size and change rates helps you find and restore a clean version.
Building a recovery plan that survives a ransomware attack? Talk to BeStarHost about dedicated infrastructure →
