A typical situation
A small business has five workstations, a file server or NAS and a “backup”: a NAS in the same office that copies files from the server at night, plus a cloud folder. One morning files have been renamed and cannot be read, and a ransom demand is on the screen. The boss wants to restore the backup. Then it turns out that the backup NAS was reachable on the same network with the same administrator account, was encrypted too, and the cloud folder had long since synced the encrypted files. The business is an invented example, the pattern is not: authorities and vendors warn about it regularly. How it can go better is shown by the real case further down.
Why attackers look for the backups first
The US authority CISA writes in its “#StopRansomware Guide” that most ransomware groups try to find reachable backups and delete or encrypt them so that recovery without paying becomes impossible. They often use stolen credentials for the backup software or vulnerabilities in unpatched backup systems. The agency recommends keeping offline and encrypted backups and regularly testing the availability and integrity of backups, and warns that automated cloud backups may not be sufficient because locally encrypted files are synced to the cloud and overwrite clean data.
A survey by Veeam among 1,000 IT leaders at businesses hit by ransomware in the previous twelve months reached the same finding in 2022: 94 per cent reported that attackers had tried to destroy the backup repositories, in 72 per cent of those cases at least partly successfully. That is a vendor survey and only a snapshot, but it fits the pattern.
Two businesses, one attack: what made the difference
In the greater Berlin area two companies were hit at the same time in 2022: the same ownership, the same office building, the same network segment. On a Friday evening the encryption started on both systems simultaneously. I was called on the Saturday morning. At one company everything was restored by Monday; the other never worked properly again. The only technical difference was the way the backup was connected to the network.
Company 1: the NAS was mapped as a network drive over SMB on every workstation, and the backup server was a domain member with domain administrator rights, because that was easier to set up. The ransomware encrypted the workstations, the mapped drives and the backup server. Data loss was total, the outage lasted several weeks, and the rebuild had to start from paper records. The ransom was paid in the end and produced a key that only partly worked.
Company 2: here Synology Active Backup for Business was running. The NAS was not in the domain, had its own local administrator account separate from Active Directory, and had no permanently mapped share – the backup ran through the Active Backup agent into a storage area that was not reachable as a network drive (reason 1 in the table below). The ransomware encrypted every workstation; the NAS was untouched. The last clean, versioned backup was from Thursday evening. Saturday for analysis, Sunday and Monday for the restore onto cleaned systems, and by Monday afternoon the business was running again. Data loss: practically none.
The conclusion from the two cases: a backup that sits in the same network segment and is reachable with the same credentials is not a backup. Synology’s immutable snapshots (WriteOnce and Immutable Snapshots, DSM 7.2, May 2023) did not yet exist at the time of the attack; what saved company 2 was keeping the domain and the share apart, and having several versions to go back to.
Seven reasons a backup does not help when it counts
| Reason | What lies behind it | Countermeasure |
|---|---|---|
| 1. Backup reachable on the same network | Drive permanently connected, share open to all, same administrator accounts | Backup target with its own credentials, not in the domain, no access from the workstation PC; the backup is pulled rather than pushed |
| 2. Synchronisation instead of backup | OneDrive, Dropbox or Drive mirror the encrypted files too | Real backups with versions that keep older states |
| 3. The backup stopped running | Error for months, storage full, password changed, nobody read the message | Notification by email to an address someone reads; monthly check date |
| 4. The wrong data | Database inconsistent during operation, program settings and licences not backed up, email and cloud data missing | Inventory of what is backed up; database with its own backup procedure |
| 5. You cannot get at the data | The key or password of the backup is on the encrypted server, the recovery program needs a licence | Keep keys separately and offline; recovery possible without the affected computer |
| 6. It takes too long | 2 TB from the cloud over the line takes days | Local copy for a quick return, off-site for the disaster; measure the time |
| 7. The backup is infected too | Malware sat in the network for weeks before striking | Several versions, older ones too; recovery in a cleaned environment |
Disconnecting the backup from the network: in practice
- One copy offline: An external drive connected only during the backup, in a cupboard or elsewhere; for businesses also tape drives or alternating two drives.
- Immutable versions: With off-site storage and current NAS systems (for example Synology) snapshots can be locked so that even an administrator cannot delete them within a set period. Synology calls the features WriteOnce and Immutable Snapshots and introduced them in DSM 7.2 (2023), so newer NAS systems have them and older installations do not. That is the most effective safeguard if the attacker gains administrator rights.
- Separate credentials and two-factor authentication for the backup system. The account for the backup must not be the same as a domain administrator account.
- Updates for the NAS and no exposure to the internet: there have been waves of attacks on NAS devices that were reachable from the internet without being patched.
- Off-site encrypted, with your own key (see The 3-2-1 backup rule explained simply, with costs).
The restore test in three levels
| Level | What | How often | Time |
|---|---|---|---|
| 1: One file | Fetch a randomly chosen file (from three weeks ago) from the backup, open it, compare | Monthly to quarterly | 10 minutes |
| 2: A folder or a machine | Restore a whole folder or a virtual machine to a different computer and check that programs start and data are correct | Yearly | 1–3 hours |
| 3: Recovery drill | Get the most important systems (file store, accounts, email, line-of-business software) running again from the backup on cleaned hardware or in the cloud, stop the clock | Every 1–2 years | Half to a whole day |
Log: what you write down
- Date, who tested, which backup (date of the version), which data or systems.
- Measured time until usable.
- What worked, what did not (missing passwords, licences, instructions).
- Measures and date for fixing them.
RPO and RTO in plain words
RPO (Recovery Point Objective) asks: How much work may be lost at most? With a nightly backup it is up to 24 hours. RTO (Recovery Time Objective) asks: How long may the business stand still at most? If you cannot be down for more than a day, the restore test has to show that recovery can be done in a day. Write both figures down: they determine how often to back up, whether a local copy is needed and what the solution may cost.
What to do when it happens
- Disconnect the network: Take the affected computers off the network (pull the cable, Wi-Fi off), do not switch them off if the IT support says otherwise. Do not click around in the ransom message.
- Keep calm and phone the contacts: Your IT support and your insurer. CERT-SE advises reporting the incident to the police at an early stage – in Sweden 114 14 is the number for contacting the police, 112 is for a crime in progress. The order of the first steps is in It has already happened: the first 60 minutes.
- Do not pay. CERT-SE is blunt about it: never pay a ransom, there is no guarantee that systems are restored, that files are decrypted or that the attacker does not come back with new demands. In the Veeam survey 24 per cent of victims said they did not get their data back despite paying, and in the case above the purchased key only partly returned the data. Do not decide that alone or under time pressure.
- Check the backup before anything is overwritten: First establish whether and where clean backups exist. Restore in a cleaned environment, not on the infected devices.
- Data protection: If personal data was affected, the breach must as a rule be reported to IMY within 72 hours of becoming aware of it (GDPR Article 33). The duty always rests with the controller, that is with the business, not with the IT provider.
- Learn: How did the attacker get in (phishing, remote access, unpatched hole)? Close that before you go back online.
What I offer
The Check plus restore test (2,500 SEK) takes about 2.5 hours, on site in Medelpad (travel included) or by remote access: I check your backup against the seven sources of error, carry out with you a restore test of levels 1 and 2 (file and folder on a separate computer), note the time and deliver a short log with an RPO/RTO assessment and a list of improvements. The price is credited against any follow-up job; the recovery drill (level 3) and the implementation I offer by time. Remote help does not qualify for RUT. Business prices exclude VAT.
Frequently asked questions
Why do ransomware attackers destroy backups?
Because working backups make the ransom unnecessary. According to CISA most groups try to find reachable backups and delete or encrypt them. That is why one copy must be disconnected from the network or immutable.
Does a cloud folder such as OneDrive protect against ransomware?
Only to a limited extent: the encrypted files are synced along and can overwrite clean data. A recycle bin and version history help only for a limited time. A real backup with several versions and an immutable copy is needed.
How often should I do a restore test?
One file monthly to quarterly, a folder or machine yearly, a recovery drill of the most important systems every one to two years, each time with timing and a log.
What do RPO and RTO mean?
RPO is the amount of data or working time you may lose at most, RTO the time the business may stand still at most. Both determine how often to back up and how fast recovery has to be.
What does the Check plus restore test cost?
2,500 SEK excl. VAT (about 2.5 hours) with a log; the price is credited against any follow-up job.
No affiliate links. Recommendations are based purely on hands-on experience.
Do you know whether your business will recover after a ransomware attack? The Check plus restore test costs 2,500 SEK (businesses excl. VAT) and is credited against any follow-up job. In English, German or Swedish, reply within 24 hours.
IT support for businesses in Sundsvall Book a free callExternal sources:
- CISA: #StopRansomware Guide (backups offline, encrypted, test; cloud-sync warning)
- CISA: Stop Ransomware
- Veeam: Ransomware Trends Report 2022 (vendor survey, 1,000 IT leaders)
- CERT-SE: Handling extortion attacks (ransomware) – never pay, 3-2-1, test restores, report early
- IMY: Personal data breaches (72 hours from becoming aware)
- EUR-Lex: GDPR (Article 33)
- Swedish Police: 114 14 for contact, 112 for a crime in progress
- Synology: Active Backup for Business (the solution in the case above)
- Synology: DSM 7.2 (22 May 2023) – WriteOnce and Immutable Snapshots