Somewhere in most organizations there's a server nobody remembers. It might be a tower in a storage closet that once ran a project database, a virtual machine left over from a migration, a cloud instance a developer spun up for a test, or a hosting account still renewing for a website that was replaced years ago. It's switched on, it's costing money, and nobody is looking after it.
These are zombie servers. They do no useful work, but they keep drawing power, consuming budget and, most dangerously, sitting unpatched on the network where an attacker might find them before you do.
A zombie server is a server that's still running but no longer does useful work. It may be physical, virtual, in the cloud or a forgotten hosting account. Zombie servers waste energy, hardware, licences and cloud spending, and because nobody maintains them, they become unpatched security risks. Finding them takes an inventory and usage data; retiring them safely takes a careful decommissioning process.
This guide explains what zombie servers are, how they appear, what they really cost, why they're a security problem, how to find them, how to retire them safely, and how to stop new ones from appearing.
What Is a Zombie Server?
A zombie server, sometimes called a comatose or orphaned server, is any server that's powered on and connected but no longer provides a service anyone needs. The term covers more than physical machines. Zombies appear wherever computing can be created and forgotten.
| Type | What it looks like | Typical origin |
| Physical server | A machine in a closet or rack with no current purpose | Replaced by a newer system but never switched off |
| Virtual machine | A VM on a host that nobody logs into | Created for a project, test or migration |
| Cloud instance | A cloud server still billing every hour | Spun up for development or a short project |
| Hosting account | A plan still renewing for a site nobody uses | A rebrand, a closed campaign or a departed staff member |
| Website or application | An old site still online on its original hosting | A redesign launched elsewhere; the old site left running |
| DNS record | A record pointing to a server or service that no longer exists | A service removed without cleaning up DNS |
The last two are the easiest to overlook and, as later sections explain, among the most dangerous.
How Servers Become Zombies
Very few zombie servers are created by carelessness alone. They usually result from reasonable decisions that were never revisited.
- Projects end, servers don't. A server built for a project keeps running after the project finishes, because shutting it down was nobody's task.
- Migrations leave the old system "just in case." After moving to a new system, the old one is kept running as a fallback, then forgotten.
- Test and development environments multiply. Each new test environment feels temporary, and few have an expiry date.
- People leave. The person who knew why a server existed moves on, and nobody else is sure whether it's safe to turn off.
- Fear of breaking something. When nobody knows what a server does, the safest-seeming choice is to leave it alone.
- No inventory. Without a current list of servers and owners, there's no way to notice when one stops being used.
Virtualization and Server Sprawl
Virtualization made it far easier to create servers, and that's made zombies more common. A physical server took a purchase order, rack space and an afternoon to install. A virtual machine takes a few clicks. The result is often called VM sprawl: dozens or hundreds of virtual machines, many of them created for a single purpose and never removed.
Sprawl has costs of its own. Idle virtual machines still reserve memory and storage on their host, which can force earlier hardware purchases. Some software is licensed per processor core or per virtual machine, so idle machines can raise licence bills. Old snapshots of virtual machines quietly consume storage. And every virtual machine is another operating system to patch.
The same discipline applies as for physical servers: every virtual machine needs an owner and a purpose, temporary machines need an expiry date, and the virtualization console should be part of every inventory review.
What Zombie Servers Really Cost
A single idle server may not seem expensive. Across an organization, and over several years, the costs add up in more ways than the electricity bill.
| Cost | How zombies add to it |
| Energy and cooling | Idle servers still draw a large share of their peak power, and every watt they use also has to be cooled |
| Hardware and space | Rack space, power capacity and hardware that could be reused or retired |
| Software licences and support | Operating system, database and application licences and support contracts renewed for machines doing nothing |
| Cloud and hosting bills | Instances, storage, IP addresses and hosting plans billed every month |
| Administration | Backups, monitoring and patching effort spent on systems that don't matter, or worse, not spent on them |
| Security exposure | Unmaintained systems that attackers can exploit |
| Data and compliance risk | Old data, sometimes including personal information, kept long after it should have been deleted |
The energy cost has an environmental dimension too. Every idle server adds to the electricity a data centre or office consumes, for no benefit. Retiring zombies is one of the simplest ways to reduce the footprint of an IT estate.
The Security Risk Nobody Is Watching
The biggest cost of a zombie server is usually not financial. It's the risk of a breach.
Zombie servers are, almost by definition, unmaintained. Nobody applies security updates, nobody reviews the accounts that can log in, and nobody notices unusual activity. They often run old operating systems and software with known vulnerabilities, use passwords set years ago, and hold data nobody remembers putting there. To an attacker scanning for weaknesses, a forgotten server is an ideal way in: once compromised, it can be used to move to more important systems, steal data or launch ransomware. Our guide to how ransomware gets into small business networks [URL pending: ransomware article] explains why unpatched, unmonitored systems are such common entry points.
Zombie websites and forgotten hosting accounts
Old websites are a particularly common kind of zombie. When a business redesigns its site, launches a campaign microsite, or changes its name, the old site is often left running on its original hosting account. It keeps renewing, but nobody updates its content management system, plugins or themes.
Forgotten websites are attractive targets. Attackers exploit outdated plugins to plant malware, host phishing pages, send spam or redirect visitors to scams, all under a domain that still carries the business's name. The business may only find out when customers complain, a browser starts showing warnings, or email from the domain starts landing in spam.
Forgotten domains, mailboxes and accounts
Zombies aren't limited to servers. Several related leftovers create similar risks:
- Expired domains. A domain the business once used, perhaps for a product or campaign, may expire and be registered by someone else, who can then receive email sent to it or put up content that trades on the old name. Our guide to registering and protecting a Canadian business domain covers how to keep domains from lapsing.
- Old mailboxes and forwarding rules. Mailboxes of former staff, or forwarding rules that send business email to personal addresses, can leak information long after anyone remembers them.
- Unused administrator accounts. Accounts on hosting control panels, content management systems and software-as-a-service tools that belong to former staff or contractors are a common way in for attackers.
Review these alongside servers. They're part of the same estate.
Dangling DNS and subdomain takeover
A subtler zombie is a DNS record that points to something that no longer exists. Suppose a subdomain such as promo.example.ca was pointed at a third-party service for a campaign, and the campaign account was later closed without deleting the DNS record. In some cases, an attacker can claim the same resource on that service and take control of what the subdomain displays, without ever touching the business's own systems. This is known as subdomain takeover.
A subdomain taken over this way can host convincing phishing pages, steal cookies scoped to the main domain, or damage the business's reputation. The fix is straightforward: when a service is removed, remove the DNS records that point to it, and audit DNS regularly for records that lead nowhere.
How to Find Zombie Servers
Finding zombies combines an inventory with evidence of whether each system is actually used.
Build or refresh the inventory. List every physical server, virtual machine, cloud account, hosting account, website and domain the organization has. Check purchase records, cloud and hosting invoices, virtualization consoles, network scans and DNS zones, since no single source is complete.
Assign or trace an owner. For each system, record who is responsible. Where nobody claims a system, it goes on the suspect list.
Gather usage evidence over a meaningful period. A few weeks of data usually shows whether a system is in use. Look at:
- Processor and memory use beyond background activity
- Network traffic, especially inbound requests from real users or other systems
- Logins by people, not just automated processes
- Disk activity and changes to data
- Application or web server logs showing real requests
Check the money. Cloud and hosting bills, tagged by project or owner, quickly show resources that nobody recognizes.
Audit DNS. Review every DNS record for each domain and confirm it points to something that exists and is still needed.
Signs a server may be a zombie
- No logins by a person for months
- Network traffic limited to updates, monitoring or backups
- Processor use flat at a low level around the clock
- No known owner, or an owner who has left
- Named after a project, campaign or person that no longer exists
- Running an operating system or application version that's long out of support
- Not in any current documentation
None of these proves a server is unused. Together, they justify a closer look.
How to Retire a Zombie Server Safely
The fear of breaking something keeps zombies alive. A careful process removes that fear.
- Confirm ownership and purpose. Ask the presumed owner, the team named on the server and anyone whose systems connect to it. Record what you learn.
- Map dependencies. Check whether other systems, scheduled jobs, integrations or DNS records rely on it. Logs of incoming connections are useful here.
- Back up and record retention needs. Take a final backup or snapshot, decide how long it must be kept, and note any data that's subject to retention rules.
- Announce the change. Tell anyone who might be affected when the server will be switched off, and who to contact if something breaks.
- Disconnect first, then wait. Shut the server down or isolate it from the network, but don't destroy it yet. If nobody complains during an agreed period, often two to four weeks, the server is very likely unused. This is sometimes called the "scream test."
- Remove the server from monitoring, backups, DNS, load balancers, firewall rules and documentation. Cancel the related licences, support contracts, cloud resources or hosting plans.
- Sanitize storage. Before reusing, selling or recycling hardware, securely erase or destroy its storage, following recognized media-sanitization guidance.
- Dispose responsibly. Send hardware to a certified electronics recycling program, and keep a record of what was disposed of and how its data was destroyed.
For websites and hosting accounts, the same steps apply: confirm nobody needs the site, take a final backup of files and database, redirect any useful URLs to the current site, remove the DNS records, and cancel the hosting plan.
When not to retire a server
Not every quiet server is a zombie. Some systems are idle most of the time by design:
- Seasonal systems, used only at year-end, during tax season or for an annual event
- Disaster recovery standbys, which should stay ready even though they're rarely used
- Systems under a legal hold, which must be preserved for litigation or investigation
- Archive systems that must stay accessible for regulatory reasons
The difference between a standby and a zombie is documentation. A legitimate idle system has a named owner, a written purpose and a place in the patching schedule. If those are missing, add them, or retire the system.
Making the Business Case
Retiring zombies takes staff time, so it helps to show what the effort is worth. A simple estimate covers:
- For physical servers, multiply each machine's typical power draw by the hours it runs in a year and your electricity rate. Add an allowance for cooling, which can be substantial in a server room.
- Licences and support. List the operating system, database, backup and application licences, and support contracts, attached to each zombie.
- Cloud and hosting. Total the monthly charges for zombie instances, storage, IP addresses and hosting plans, and multiply by twelve.
- Note any hardware purchases that retiring zombies would postpone, and any rack or room space freed.
- Staff time. Estimate the time spent backing up, monitoring and patching systems that don't need to exist.
- A breach through an unpatched zombie can cost far more than everything above combined. You don't need a precise figure; it's enough to show that the organization is carrying unmanaged risk.
Presented together, these usually make a clear case, especially when the first review finds systems nobody can explain.
Data Disposal and Privacy in Canada
Zombie servers often hold personal information: old customer databases, email archives, HR files or website contact-form submissions. Retiring them is a chance to deal with that data properly.
The federal private-sector privacy law, the Personal Information Protection and Electronic Documents Act (PIPEDA), expects organizations to keep personal information only as long as needed and to take care when disposing of or destroying it, so that unauthorized parties can't access it. Keeping years of personal data on an unmonitored server conflicts with both expectations, and a breach of that server could create reporting obligations.
The Canadian Centre for Cyber Security publishes guidance on sanitizing IT media, covering methods appropriate to different types of storage and sensitivity levels. Following recognized guidance, and recording what was done, gives you evidence that disposal was handled properly.
For the hardware itself, most provinces have regulated electronics recycling programs that accept servers and computer equipment. Using them keeps electronic waste out of landfill and gives you documentation of disposal.
Preventing the Next Zombie
Clearing out zombies once helps. Stopping them coming back requires a few habits.
- Every system has an owner. No server, cloud resource, hosting account or website is created without a named person responsible for it.
- Tag and label everything with owner, project and purpose, so anyone can see what a resource is for.
- Give temporary systems an expiry date. Test environments, campaign sites and project servers should be created with a planned end date and a reminder before it.
- Build retirement into project closeout. When a project or migration ends, retiring its systems is part of the checklist, not an afterthought.
- Include systems in staff offboarding. When someone leaves, review the systems and accounts they owned and reassign or retire them.
- Review regularly. A quarterly review of inventory, usage, bills and DNS catches zombies early.
- Fewer, better-managed systems are easier to keep track of than many scattered ones.
Zombies in the Cloud
Cloud computing makes zombie servers easier to create and easier to forget. Anyone with access can start an instance in seconds, and the cost appears only on a monthly bill that may go to someone else.
Common cloud zombies include instances left running after testing, development environments running around the clock when they're used only during working hours, unattached storage volumes and old snapshots, reserved IP addresses no longer assigned to anything, and load balancers or databases left behind after an application is removed.
The same practices apply: tag every resource with an owner, set budgets and alerts, review bills monthly, and schedule non-production environments to shut down outside working hours. Our guide to trimming what you pay to store data in the cloud [URL pending: cost article] covers storage-related waste, such as orphaned volumes and snapshot sprawl, in more detail.
Consolidation and the Hosting Option
Zombie servers are often a symptom of too many small, scattered systems: a server in the office for one application, a cloud instance for another, and websites spread across several hosting accounts opened over the years. Consolidating them reduces both cost and the number of places a zombie can hide.
For many small and mid-sized organizations, a practical step is to bring websites and lightweight applications together with one hosting provider, under one account with a clear owner and renewal schedule. Depending on the workload, that might mean shared hosting for simpler business sites, a managed WordPress plan, or a VPS that can replace an ageing office server. Moving workloads off an old in-office machine also removes the power, cooling and hardware-replacement burden from the office.
4GoodHosting hosts customer sites and applications in Canadian data centres in Vancouver and Toronto, which keeps consolidated workloads under Canadian jurisdiction. Whichever provider you choose, keep the inventory current and the ownership clear, or the zombies will simply move with you.
Three Scenarios
A Winnipeg accounting firm finds three servers in its storage room during an office move. One runs the current practice software. Another was replaced two years earlier but still holds client files from a previous system. The third has no known purpose. After confirming with staff and checking logs, the firm backs up the old client data under its retention policy, runs a four-week scream test, and then has both unused servers securely wiped and recycled. Its electricity use and the noise in the storage room both drop noticeably.
A digital agency reviews its hosting accounts and finds more than a dozen campaign microsites and old client sites still online, several running outdated plugins. Two have already been compromised and are serving spam pages. The agency takes final backups, redirects useful URLs to current sites, removes the old DNS records, and cancels the accounts. It now gives every campaign site an expiry date when it's launched.
A software startup sees its cloud bill rise every month. Tagging shows that most of the growth comes from test environments created for features that have long since shipped. The team deletes environments with no recent activity, schedules the rest to shut down overnight and on weekends, and requires an owner and expiry tag on every new resource.
A Quarterly Zombie Server Checklist
- Compare the inventory against invoices, cloud consoles, virtualization hosts and network scans.
- Confirm every server, cloud resource, hosting account and website has a current owner.
- Review usage data and flag systems with no meaningful activity.
- Audit DNS for records pointing to removed services.
- Check that every live system is receiving security updates.
- Review cloud and hosting bills for unrecognized charges.
- Retire flagged systems using the safe decommissioning process.
- Confirm retired hardware was sanitized and recycled, with records kept.
FAQ
What is a zombie server?
A zombie server, also called a comatose or orphaned server, is a server that's still powered on and connected but no longer does useful work. It can be a physical machine, a virtual machine, a cloud instance, a hosting account or an old website. It keeps consuming power, budget and attention, and because nobody maintains it, it becomes a security risk.
Why are zombie servers a security risk?
Nobody patches them, reviews their accounts or watches their activity. They often run outdated software with known vulnerabilities and hold forgotten data. Attackers look for exactly these systems, because a compromised zombie can be used to move into more important systems, steal data or launch ransomware.
How do I find zombie servers?
Build an inventory from purchase records, invoices, cloud consoles, virtualization hosts, network scans and DNS zones. Assign an owner to each system, then gather usage data over a few weeks: processor use, network traffic, human logins, disk activity and application logs. Systems with no owner and no meaningful activity are likely zombies.
What is a scream test?
A scream test means disconnecting or shutting down a server you believe is unused, without deleting it, and waiting a set period, often two to four weeks, to see whether anyone notices. If nobody complains, the server is very likely safe to decommission. Always take a backup first.
How do I safely decommission a server?
Confirm the owner and purpose, map dependencies, take a final backup and note retention needs, announce the change, disconnect and wait, then remove the server from DNS, monitoring, backups and firewall rules. Cancel related licences and services, securely sanitize the storage, and recycle the hardware through a certified program.
What is subdomain takeover?
Subdomain takeover happens when a DNS record still points to a third-party service that was removed. In some cases an attacker can claim that resource on the service and control what the subdomain displays. The fix is to remove DNS records when the services they point to are removed, and to audit DNS regularly.
Do zombie servers exist in the cloud?
Yes, and they're often harder to spot. Instances left running after tests, development environments running around the clock, unattached storage, old snapshots, unused IP addresses and leftover databases all keep billing. Tagging every resource with an owner, setting budget alerts and reviewing bills monthly helps catch them.
Key Takeaways
- Zombie servers are running systems that no longer do useful work: physical, virtual, cloud, hosting accounts, old websites and dangling DNS records.
- They cost energy, hardware, licences, cloud fees and admin time, and their biggest cost is security risk.
- Forgotten websites and DNS records pointing to removed services are among the most commonly exploited zombies.
- Find zombies with an inventory, clear ownership and a few weeks of usage data.
- Retire them safely: confirm, map dependencies, back up, announce, disconnect and wait, then decommission, sanitize and recycle.
- Prevent new zombies with owners, tags, expiry dates, project closeout steps, offboarding reviews and quarterly checks.
Conclusion
Zombie servers are an ordinary problem with serious consequences. They appear when projects end, people leave and migrations finish, and they survive because nobody is sure it's safe to turn them off. Meanwhile, they draw power, renew licences and hosting plans, and wait unpatched for an attacker to find them.
The fix is not complicated. Know what you run and who owns it. Use evidence, not guesswork, to decide what's still needed. Retire what isn't through a careful process that protects data and avoids surprises. And build habits, such as expiry dates, closeout steps and regular reviews, that stop new zombie servers from taking their place.
The reward is a smaller, cheaper, more secure environment where every system has a purpose and someone looking after it.
If your websites and applications are scattered across old servers and forgotten hosting accounts, consolidation is the simplest way to keep them under control. With 4GoodHosting you can bring your websites and applications under one account, hosted in Canadian data centres you can name, with one renewal schedule and a clear owner. Talk to the team about moving workloads off ageing office hardware and retiring the accounts you no longer need.








