A backup is worth so much how long was the last successful recovery test
We design and maintain servers, virtualisation and backups. With a reproduction test that ends with a protocol, not an assurance that everything works.
When it's worth tackling
If you recognise two or three of these signs, a free IT review will show you where to start.
- The backup exists, but no one remembers when the data was last restored from it.
- All copies are on the same device, in the same room as the server.
- The key system runs on a single computer that sits under the desk.
- The server is more than five years old and is no longer supported by the manufacturer.
- There is no answer to the question of how many hours a company can be off work without serious consequences.
- Data is in several places at once and no one knows which version is current.
- Copies are made "to an external drive", and the disk is connected by the person who remembers.
What we do as part of this service.
- 1
Environment review and design
Inventory of servers, arrays, virtualisation and copies. We determine how much time the company can stop working and how much data it can lose, because the rest of the project depends on these two numbers.
- 2
Servers and virtualisation
Selection, installation and configuration of the server and virtualisation layer. Separating services into virtual machines so that updating one thing does not risk the others.
- 3
File server and permissions
Common space for documents with permissions reflecting the company structure. With file version history that saves after a simple human mistake.
- 4
Triplicate backups
Two copies on different media and one off-site. Offsite copy is the only one that helps with fire, flood and theft.
- 5
Backup of the entire system, not just files
Restore the system with applications and settings, also to other hardware. A copy of the documents themselves does not reduce downtime, because the system has to be built from scratch.
- 6
Playback Test
Cyclic recovery of the selected system in a test environment, with time measurement and protocol. Included in the PROFESSIONAL IT package every quarter.
- 7
Monitoring and alerts
Disk status, free space, temperature, copy job status. The alert comes to us before the drive fails, not after the fact.
- 8
Disaster recovery plan
Document with system recovery order, responsible persons and access data stored securely. Without this, playback starts with searching for passwords.
What you get in writing.
Server environment documentation
Scheme, addressing, list of virtual machines with their roles, dependencies between services and schedule of copy tasks.
Backup Policy
What is copied, how often, where it is stored, how long it is stored and who has access. On one page, in a language the board understands.
Recovery Test Protocol
Date, range, measured playback time and result. A document that can be shown to an insurer or contractor.
Disaster recovery plan
Restore order, decision points, contacts and where access data is stored. Updated whenever there is a significant change in the environment.
Who is responsible for what.
Allowable downtime is a business decision, not a technical one. We can show you how much it costs to reduce it from eight hours to two, but the choice is yours.
On the NexaIT side
- server environment design and backup layer
- installation, configuration and maintenance of covered servers and virtualisation
- configuration of copy tasks and monitoring their execution
- cyclical copy verification and recovery test according to the scope of the package
- updates to the virtualisation layer and supported systems
- maintaining environment documentation and recovery plan
- conducting disaster recovery
On your side
- decision on permissible downtime and permissible data loss, because it is a business decision
- budget for hardware, licences and off-site copy space
- ensuring conditions in the server room: power supply, cooling, access control
- maintaining an Internet connection with parameters enabling sending a copy outside the premises
- indication of critical data and systems, because not everything needs to be copied with the same frequency
- consent to the service window needed for updates and testing
- decision to initiate a disaster recovery plan
What this service doesn't include.
Clear scope boundaries are part of a good contract. We talk about them before you sign, not when the first invoice arrives.
We select equipment and licences together with you and order on your behalf, if you prefer. We do not hide these costs in the price of the service, because then it would be impossible to compare offers.
- server equipment, arrays and disks, which we invoice separately as a pass-through cost
- virtualisation licences, operating systems and backup software
- off-site copy space, which is a recurring cost with the third-party vendor
- data recovery from physically damaged media, which is a laboratory service
- guarantees the restoration of data that has never been backed up
- reconstruction of the electrical and cooling installations in the server room
- support for an industry application for which the manufacturer does not provide recovery documentation
When it's a good fit, and when it isn't
It makes sense if
- The company has a system whose downtime stops the sale, production or release of goods.
- Data must remain on your infrastructure for contractual or industry reasons.
- You want to know the measured playback time, not hear that there is a copy.
- The environment is growing and maintaining additional computers as servers no longer makes sense.
We don't recommend it if
- All data is already in the cloud and there is nothing local to secure. Then the conversation is more about a copy of data from the cloud.
- Are you looking for the cheapest server without maintenance service. Unattended equipment itself is a cost, not a solution.
- There are no conditions for a server room and no budget to create one.
- The company is not ready for service windows. Without them, the environment cannot be updated safely.
What affects the price
- allowable downtime and allowable data loss, because they enforce the architecture
- number of servers, virtual machines and data size to be copied
- whether high availability is needed, i.e. operation after the failure of one node
- number of locations and method of sending copies offsite
- required backup storage period, because this is directly converted into space
- number of industry systems with non-standard recovery procedure
- input state: migrating from a live environment is more expensive than building from scratch
Questions before you decide.
How often do you test playbacks?
In the START IT package, we verify the copying every month, but without full restoration. In BUSINESS IT, verification is automatic. In PROFESSIONAL IT and ENTERPRISE IT, we perform a recovery test every quarter and end with a protocol with a date and measured time. In addition to the package, the test can be ordered as a separate service.
Does the data have to leave our company?
One copy should be kept off-site because it is the only layer of protection against fire, flooding and theft. This can be done in several ways: space at the supplier, a device at a second location of your company or a medium replaced on a schedule. If the data cannot leave your infrastructure, we choose the second or third option.
Server in the company or in the cloud?
It depends on what the system does and who uses it. The application used by production in the hall should usually run locally because it cannot depend on the connection. Mail and files usually work better in the cloud. The most common result that makes sense is a mixed model and we do not sell one answer to every question.
What about a server that's eight years old and still running?
It works until it stops, and then the question is how quickly it can be replaced. As a matter of principle, we do not encourage exchanges. We check whether there is a playback path to other hardware and how long it takes. If recovery is feasible within an acceptable time frame, older hardware may still be operational.
From our projects, without client names.
All case studiesRelated
Free · 60 minutes online · no obligation
An IT review, not a sales call.
- You talk to an engineerOnline, by video call. Not with a salesperson. We don't install or change anything.
- We check 8 areasBackups, access, network, email, server, licences, protection and KSeF readiness.
- You get a scorecardThree priorities on one page, emailed after the meeting. Yours to use however you like.
We don't use a contact form. We answer the phone and reply to emails.