Ransomware in a company in Algeria (2027): what to do in the first hour when every file is encrypted, and the five barriers against being hacked

“This morning not a single file opens. They all have a strange extension, and a message on the screen demands a ransom to give them back. The server, the shared folder, the accounts: everything. We have a backup on a disk, but it was plugged in. What do we do, right now?” The first hour is not for repairing: it is for stopping the spread and keeping what is still healthy. And the attack did not start this morning: it started three weeks earlier, with a password, an open door and a network with no partitions.

Symloop13 min read
Ransomware in a company in Algeria: first hour, barriersSeven actions in order, the attack chain, five barriers, an out-of-reach copy, the recoveryRANSOMWARE IN A COMPANY · ALGERIA · 2027Ransomware in a companyin Algeria: first hour, barriersSeven actions in order, the attack chain, five barriers, anout-of-reach copy, the recoveryEVERY FILE ENCRYPTEDa message demands a ransomAN OUT-OF-REACH COPYand a recovery already triedsymloop.com

Short answer. Facing ransomware, the first hour comes down to seven actions, in order: cut the internet; isolate the affected machines from the network, without switching them off or restarting them; put out of reach the backups that are still healthy by unplugging them; alert the manager and the IT provider by phone; note the time, the machines affected and every action, and photograph the message; pay nothing and reply to nothing in a hurry; decide who leads and what restarts first. Recovery then happens on clean machines, after finding how the attacker got in and changing every access. And so that it does not happen, or so that the company restarts in two days, five barriers: accounts, doors, partitions, out-of-reach backups, detection.

This guide is for the manager living that morning, the one who has watched a peer live it, the company whose only backup is a disk plugged into the server, and the founder living in France who wants to know what their Algerian company would do in that case. The first-hour actions follow official public guidance, from the French victim-assistance platform and from the US cybersecurity agency, which converge: isolate, do not switch off affected machines, do not pay, keep the evidence, reinstall, change all passwords, bring services back gradually. The attack, its dates and the company in the example are constructed from public descriptions of these attacks, not surveyed at a client's. Filing a complaint, mandatory reporting, insurance and the question of payment are matters for your adviser and the authorities; this guide stops at the technical actions.

Six sentences from that morning, and the right answer

Each row is a natural reflex, and what to do instead.

Figure 1 · Six sentences heard on the morning the files no longer open, the reflex each one triggers and what to do instead. The first three play out in the first hour: isolate without switching off, do not pay in a hurry, keep the backup away. The next three in the days that follow: measure the extent, find the way in, put up the barriers.

“Switch everything off!” The reflex is to cut the power; official guidance asks for the opposite: cut the connections, the internet first, then the network of the affected machines, and do not switch them off, because the traces useful to the investigation and sometimes to recovery are in memory. A shutdown is considered only if encryption is visibly still running and the machine cannot be isolated. “Let's pay and be done with it”: paying guarantees neither the key nor that the copied data will not be published; official guidance is not to pay, and it is a decision that is taken neither within the hour nor alone. “Plug the backup in, quick”: as long as the attacker or their program is active, anything plugged in is encrypted in turn; the healthy backup stays unplugged until a clean machine is ready.

“It's only one PC”: the message appears on one PC, but encryption has most often hit the server and the shares first; the extent is checked before anyone is allowed to resume. “We reinstall and it's over”: if nobody knows how they got in, they will come back the same way, with the same passwords; the way in is found, every access is changed, then machines are reinstalled. “It won't happen to us again”: with no barrier in place, nothing has changed; the last step of recovery is a written list of what was missing, with a date beside each line. There is also a public project, No More Ransom, supported by Europol and the Dutch police, which provides free decryption tools for some ransomware families: it is worth a look before any other decision, without expecting a miracle.

01

The first hour: seven actions in order, five things not to do

The figure places the seven actions on the hour that follows the discovery.

Figure 2 · The first hour, from 08:05 to 09:05, following official public guidance: cut the internet, isolate the affected machines without switching them off, put the still-healthy backups out of reach, alert by phone, note what is done, decide who leads. On the right, five things not to do. Official guidance asks not to switch off affected machines, to keep the traces in memory, and to consider a shutdown only if encryption is still running and they cannot be isolated. Times and machines are a constructed example.

Cut the internet: at 08:07, two minutes after noticing, the operator box and the backup link are unplugged; the attacker loses control and can no longer steer the encryption or take out more data. Isolate the affected machines: network cable unplugged, Wi-Fi off, tunnels between sites closed so the branch is not reached; the machines stay on, and those that were off are not switched on. Put the backups out of reach: the backup disk or NAS not yet affected is unplugged and set aside; it is the most valuable asset of the day.

Alert: the manager and the provider are called from a printed list, because the directory, the email and sometimes the phone system are on the affected machines. Note and keep: a photo of the message, the time, the list of machines, a log of every action; these serve the investigation, the insurance and the search for the way in. Decide: one person leads; the team is told to reconnect nothing; customers waiting for a delivery or an invoice are told by phone; what restarts first is chosen. A till that works offline and email hosted elsewhere make it possible to keep selling and answering while the rest is stopped.

«The first hour is not for repairing: it is for stopping the spread, keeping what is still healthy and noting what happened.»
02

Twenty-one days earlier: the attack chain, and the five barriers

The figure goes back three weeks and shows, under each link, what would have broken it.

Figure 3 · The same attack seen three weeks earlier: a password typed on a fake page, an open mailbox, a remote desktop reachable from the internet, an administrator account, the backups reached, then encryption on a Friday night. Under each of the first five links, the barrier that would have broken it. The fifth, an out-of-reach copy whose restore has been tried, is the one that decides whether the company restarts in two days or has nothing left to restore. Dates and steps are a constructed example, following public descriptions of these attacks.

The phishing email and the open mailbox: ransomware most often comes in through an attachment or a link in an email, or through a flaw in an exposed service; here, an employee types their password on a fake page, and the attacker opens the mailbox, reads the exchanges and finds other access details. Two-step sign-in makes the stolen password insufficient; an alert on an unusual sign-in gets the account blocked the same day, as in the guide on email. The door on the internet: a remote desktop left open “to fix something”, with the same password; no service should be reachable from the internet, and updates close known flaws.

The administrator account: on a network with no partitions where everyone is administrator of their own PC, the attacker moves from machine to machine; separated networks, reduced rights and an administration account distinct from the everyday one slow them down and make them visible. The backups: before encrypting, they look for the backups and destroy them, because they know backups are what make the ransom pointless; a disk permanently plugged in or a NAS on the same network are not backups that day. Encryption comes last, on a weekend night, often after copying data in order to threaten to publish it. Each of the five barriers would have been enough to break the chain; the fifth is the one still useful when the first four have given way.

03

Five barriers, in layers

Backups are the accented layer: the only one still useful when all the others have given way.

Figure 4 · The accented layer is backups: it is the only one still useful when all the others have given way, provided one copy is out of reach and the restore has been tried. Accounts, doors and partitions, above, make the attack harder and slower; detection and the plan, below, win the minutes of the first hour.

Accounts. Two-step sign-in on email, remote access and administration; one account per person, closed on the day they leave; a separate administration account, used only to administer, as in the guide on roles. Doors. No remote desktop and no administration interface open on the internet; access goes through a tunnel; the firewall, servers, workstations and NAS are updated, because attacks exploit flaws the vendors have already fixed. Partitions. Separated networks stop a PC from reaching the cameras, the backups or administration; shares are open only to those who need them; a workstation without administrator rights limits what a program launched by mistake can do.

Backups. The usual rule is three copies of the data, on two different media, one of them off site; against ransomware an out-of-reach copy is added, offline or immutable, which even an administrator account cannot delete, and the restore is tried, as in the guide on backups. A backup from which a whole server has never been restored is a hypothesis. Detection and plan. An alert when an account signs in from an unusual place, when a PC changes thousands of files in a few minutes, when a backup fails; a call list and the seven actions printed and pinned up; a one-hour exercise each year, where things are unplugged for real. None of these barriers is a miracle product: they are settings, habits and a copy stored elsewhere.

«A backup from which a whole server has never been restored is a hypothesis. The out-of-reach copy is the only barrier still useful when all the others have given way.»
04

The five barriers in 2027, in weeks

Four weeks in the example, the first with nothing changed.

Figure 5 · Four weeks in the example, the first with nothing changed: accounts are reviewed, what is open on the internet is listed and, above all, a restore is tried. The accented week is partitions and backups, because it decides what is left after an attack. The exercise in the last week takes one hour and is repeated every year.

Week 1, stock-take: list of accounts and of those who are administrators, services reachable from the internet, shares and their rights, existing backups and, above all, a trial restore of a folder and then of a server; this is often when it turns out the backup cannot be restored. Week 2, accounts and doors: two-step sign-in switched on wherever it exists, separate administration accounts, old accounts closed, open ports closed, remote access moved onto a tunnel, updates applied.

Week 3, partitions and backups: separated networks, rights on shares and workstations reduced, out-of-reach copy put in place, full restore tried and timed; the measured time becomes the recovery time the company knows. Week 4, detection and exercise: alerts set, call list printed, first-hour exercise with the team, a one-page report. A ten-workstation company with no server goes faster; several sites, an old server or software that demands administrator rights take longer.

05

Recovery, in six steps

Restoring onto clean machines is the accented step: healthy data is not put back where the attacker may still be.

  • What the manager sees: the date of the last restore trial and how long it took, accounts without two-step sign-in, services open on the internet, the week's alerts.
  • What the company no longer does: leave a backup disk plugged in, share an administrator password, open a remote desktop to the internet, discover on the day of the outage that the backup cannot be restored.
  • What is measured: timed recovery time, age of the last out-of-reach copy, share of accounts with two-step sign-in, number of services open on the internet, which must stay at zero.
Figure 6 · The accented step is restoring onto clean machines: healthy data is not put back on a machine where the attacker may still be. Before that, isolation is held, the way in is found and every access is changed from a safe machine; afterwards, services reopen in order of priority, and what was missing is written down with a date for each barrier.

Hold the isolation: nothing reconnects before it has been checked; the team works with what was not affected, phone, hosted email, offline till, paper. Find the way in: how they got in, which accounts they used, which machines they touched, since when; this work is done with a professional, on the machines left switched on. Change every access: all passwords, email, administration, bank, suppliers, are changed from a safe machine, because the previous ones are in the attacker's hands.

Restore onto clean: affected servers and workstations are reinstalled, not “cleaned”; the backup is verified on an isolated machine before being restored, and a copy from before the attacker got in is used if one exists. Reopen by priority: the till and invoicing first, then stock, then the rest; each reopened service is watched closely, and the return to service is gradual. Write what was missing: the list of absent barriers, with an owner and a date; it is the only document from the crisis that is still useful the following year. In the example, with an out-of-reach copy, the company restarts in two days; that time depends on the volume, the hardware available and what had been tried beforehand.

What should we do first when every file is encrypted?
Cut the internet, then unplug the affected machines from the network without switching them off, then unplug the backups that are still healthy. Only then call the manager and the IT provider, photograph the message and note the time and the machines affected. Do not pay and do not reply to the message in a hurry.
Should we switch the computers off?
No, with one exception. Official guidance asks not to switch off affected machines, because useful traces are in memory, and not to switch on those that were off. They are isolated from the network. A shutdown is considered only if encryption is still running and they cannot be disconnected.
Should we pay the ransom?
Official guidance is not to pay: nothing guarantees the key or that the copied data will not be published, and payment funds the next attack. It is a decision for the manager, with their adviser and the authorities, never in the first hour. This guide does not settle that question; it makes sure it does not arise, thanks to an out-of-reach copy.
Can the files be recovered without paying?
From the backup, if one copy stayed out of reach: that is the normal case. Without a backup, a free decryption tool sometimes exists for the ransomware family involved; the public No More Ransom project, supported by Europol and the Dutch police, lists them. Sometimes nothing can be recovered. In every case, a copy of the encrypted files is kept: a tool may appear later.
How did the hacker get in?
Most often through a phishing email, an attachment or a link to a fake sign-in page, through a reused password, or through a service open on the internet and not updated, such as a remote desktop. They are usually present several days before encryption, long enough to find an administrator account and the backups. Finding the way in is part of recovery: without it, they come back.
Is an antivirus enough?
No. It stops some known programs, not an attacker using a real account and the company's own administration tools. What protects is two-step sign-in, no door open on the internet, separated networks, updates applied, and an out-of-reach backup copy whose restore has been tried.
Our backup is on a disk or NAS that stays plugged in: is that enough?
No. A disk permanently plugged in or a NAS reachable from the workstations is encrypted with the rest; the attacker even looks for the backups first. At least one out-of-reach copy is needed: a disk unplugged and stored elsewhere, or an immutable copy with a hosting provider, which nobody can delete for a set period. And a whole-server restore must have been tried.
How long does it take to restart after an attack?
It depends on what had been prepared. With an out-of-reach copy and a restore already tried, a mid-sized company restarts in a few days; two in the example. Without a healthy copy, recovery is counted in weeks, rebuilding data from paper, emails and partners. Your company's real time is measured by a restore trial, not by an estimate.
Tell us where your backups are, who is an administrator and what is open on the internet: we come back with a restore trial and the list of missing barriers

A company that has just been attacked, one that has watched a peer go through it, or a founder in France who wants to know whether their Algerian company would restart: we take stock, put up the five barriers, set up the out-of-reach copy and run the first-hour exercise with you

A one-week stock-take, a timed restore, a sheet pinned up with the seven actions. If an attack is under way, the first-hour actions are those in this guide, and the investigation is carried out with the authorities and your adviser. We work from Algiers and Sétif, in French, Arabic and English.