DATA SHEET AT-02-S · SERVER IN OPERATIONREV. 2026-A
Windows servers for your business
In most businesses the shared folders, the industry software and the sign-in on every PC depend on the server. When it stops, the office stops. We plan, replace and look after Windows servers for businesses in the Vogtland so that exactly this does not happen on a Monday morning.
You get a server that suits your work (and not the catalogue page), plus a system file that records what runs and how it is set up. The access credentials stay with you. You can reach us Mon–Fri 8:00–17:00 on +49 3744 4399760; the initial consultation is free.
01 / WHY A SERVER OF YOUR OWN
When the server stops, the office stops.
The server is the device nobody thinks about as long as it runs. It holds the shared folders, the database of your stock and order system or your practice software, and the list of who may sign in at which PC. If it goes down one morning, nobody writes an invoice, and half the office ends up tidying their desks.
Whether you need a server of your own is not decided by the number of staff. A small trade business whose industry software needs its database on the premises needs one. A considerably larger office that only stores files and already works with Microsoft 365 may get by without. We will tell you so.
That is why we start with your work and not with the technology. What has to be running first thing in the morning? What can be missing for half a day, and what absolutely cannot? That tells us which services you need, how powerful the machine must be and what the backup looks like. We never look at the server in isolation, but together with workstations and network, as in our IT support. The page Servers & virtualisation gives an overview of both topics.
02 / WHAT THE SERVER DOES FOR YOU
Four jobs your working day depends on.
Shared folders: everyone works on the same files.
Quotes, drawings, scans, the accounts filing: the file service makes them available as a drive at every workstation. If it fails, nobody can reach their documents (and anyone with a copy on their desktop is from then on working with a second version of the truth). We set up the filing to suit your departments. When changing servers we also carry over drive letters and paths, otherwise the shortcuts point to nothing the next morning.
Industry software: stock and orders, time recording, practice software.
Programs that several workstations use at the same time need a server that holds their database. When it is down, nobody writes an invoice any more. Which Windows version and which specification the software needs is set by its manufacturer. We provide exactly that foundation and coordinate server updates with the manufacturer, so that the program does not suddenly stop working after a maintenance evening.
Sign-in and permissions: who gets in and what they see.
Active Directory (AD for short) is the central user and permissions management of Windows. It records who belongs to the business, which password applies and who is in which group. The server that keeps this list is called the domain controller. If it cannot be reached, signing in at the PCs only works in a limited way, and nobody gets to the drives. In everyday terms: a new employee gets an account and can start work. Someone who leaves is locked out in one place and is then out everywhere.
The internal address book: DNS.
DNS translates names into addresses, inside your own premises too. The PCs ask it where to find the sign-in server and the drives. If it is set up wrongly, it looks like a mystery from outside: the internet works, but signing in takes minutes and the drives are missing. We look for such faults at the affected workstation, on the path it actually takes. And after every change to the network we test the internal services, not just a web page.
How many servers these jobs are spread across is something we decide with you. The system file then also records what depends on what (the industry software needs the sign-in, the sign-in needs DNS), so that when something goes wrong nobody looks in the wrong place. If the problem is more likely to lie with the switch, router or Wi-Fi, see networks.
03 / SURVEY
What runs today, and who still knows?
Before we change anything, we record what is there. We do that on site, and you do not have to gather any documents: we look at the servers, the Windows versions, the programs on them and which workstations, printers and machines use them.
The interesting part is what nobody keeps in mind any more. The export that sends figures to another system every night. The shortcut that still points to the name of the previous server. The service that runs under the account of an employee who left the business long ago. As long as the old server runs, none of this shows. The day after the move, it does.
This works even if there are no documents and the former service provider can no longer be reached. We gain access in an orderly way and write down what we find. What cannot be clarified goes on the list as an open point, not into the quote as a guess. In the end you have (often for the first time) a sheet that says what your server actually does.
04 / THE RIGHT SIZE
How big does the server need to be?
What runs at the same time?
An office that opens files all day puts a different load on a server than a stock and order system that runs the month-end close in the evening. We ask about your busy periods and take the software manufacturers' requirements as the lower limit. We plan processor, memory, drives and network together. A fast processor is no use if the disks cannot keep up.
Where does the waiting come from today?
If the stock and order system is sluggish in the morning, it is not always the processor. Often it is the drives or the cable between workstation and server. That is why we measure on the existing system before ordering anything. Then you know which component is the bottleneck, and you do not pay for performance in the wrong place.
How much headroom, and for what?
Headroom costs money, so it gets a reason: data growth over the next few years, new workstations, the second specialist application you are already thinking about. The quote states which assumptions we start from. If something bigger comes later, we recalculate beforehand.
That leaves the question of the case. A tower stands on its own and fits into a side room. A rack server belongs in a server cabinet, where the switch and patch panel are usually already mounted. The form factor says nothing about performance. We decide by your room: does the device fit, can you reach it to swap a disk, and does the fan bother someone making calls right next to it?
And: does every job really need a device of its own? Several servers can run as virtual machines on one. When that pays off is explained on the page about virtualisation with Hyper-V or VMware. If you wish, we order the server for you as well; how that works is described under hardware.
05 / DISKS, POWER, LOCATION
A disk fails, the power goes. What then?
Drives break; with servers that is expected. That is why the disks run as a group, the RAID: the data lies on several disks at once, and if one fails, the server carries on with the others. If we look after your server, it reports the faulty disk to us, and all you notice is that we arrange an appointment to swap it. The available methods are shown in Dell's overview of RAID levels.
That is not a backup. If someone accidentally deletes the folder with the quotes, it is gone from every disk in the group at the same moment. The same goes for a file that was overwritten by mistake. For that you need a backup with older versions.
For space we calculate two things separately. First: how much do you use today, and how fast is it growing? (Scans, building-site photos and CAD files grow faster than you would think.) Second: how fast do the disks have to respond? A database that a whole office works on at once needs fast drives, even if it takes up hardly any space. A server that fills up simply stops. That is why we keep an eye on the free space when we look after it.
Then the power. A UPS (uninterruptible power supply, at heart a large battery in front of the server) bridges short outages and, in longer ones, shuts the server down cleanly instead of cutting it off in the middle of writing. We size it by what is actually connected, and we test the shutdown once. If a UPS is already there, we first look at the batteries. They age even if the case still looks fine. More in the guide “Power cut at work”.
And the room: a server needs air and little dust. The cupboard next to the heating pipe is a poor place, even if the network socket happens to be there. We look at the location and tell you what is possible. Work on the electrical installation is done by your electrician; we tell them what we need.
06 / ACCOUNTS, GROUPS, FOLDERS
Who may access which folders?
You decide who may see what. Payroll is none of the apprentice's business, and the costing is not for everyone in the warehouse. We set this up so that someone can still understand it later: permissions are tied to groups (“Accounts”, “Management”, “Workshop”) and not to individual people. When someone new joins, they go into the right group and have everything they need. Nothing more.
Technically there are two locks between an account and a folder: the permissions on the share (the drive on the network) and the permissions in the server's file system. Both must allow access. That is behind the classic problem: someone is signed in, sees the drive and still gets “Access denied” on opening it. Rather than poking around in individual permissions, we follow the path from the account via the group to the folder.
In the end we test with an ordinary user account, not with the administrator. The administrator may do everything, so it always works for them. And we also try the opposite direction: can somebody from the workshop really not get into the management folder?
That leaves the question of whether Microsoft 365 makes all this unnecessary. The accounts there (Microsoft Entra ID) and the user management on your own server (Active Directory) are two different things with different jobs; Microsoft sets them side by side in its comparison of Active Directory and Entra ID. If your industry software depends on the sign-in on the premises, the server stays. We look at what depends on what in your business.
How we set up Microsoft 365 itself and move mailboxes over to it is described on the page Microsoft 365.
Incidentally, nobody gets administrator rights just because a program is acting up. First we check which permission the program is actually missing.
07 / YOUR INDUSTRY SOFTWARE
The software says what the server must be able to do.
For stock and order systems, time recording or practice software, the rule is: the manufacturer states which Windows version and which database its program may run on. We stick to that. If a program installs and starts on a newer server, that only means it starts. If something goes wrong later, the manufacturer says: not approved, no support. That is why we get the approval before we order the server.
The division of work is clear. We build and look after the server the software runs on. The manufacturer is responsible for the software itself (screens, booking logic, reports). If, when something goes wrong, it is unclear whose side the fault is on, we coordinate with them directly. You do not have to mediate between two companies.
Before a change we record everything attached to the software: the label printer in the warehouse, the card terminal, the interface to the bank, the nightly import from the web shop. If the name or address of the server changes, these are exactly the things that stop working, and without any error message.
08 / THE MOVE
The move: outside your busy periods, with a rehearsal and a way back.
A server migration is like an office move with a furniture list: what runs today, what has to come along, what can go? We have the list from the survey. From it we make a schedule that states who does what on the day, who signs it off on your side and which software manufacturer has to be reachable.
We build the new server in advance and test it while the old one keeps running. For the cut-over day we find a window with you outside your busy periods. During that time nobody may write to the old server any more (not even a nightly import), so that the latest state comes across in full. How long the pause lasts depends on the amount of data; the planned duration is in the schedule. If you would like a weekend, that is possible too. The surcharge from the price list applies, and it is stated in the quote beforehand.
After that we check on the new server what your people will do first the next morning: open drives with an ordinary user account, create a document in the stock and order system, print, trigger the export. Only when that works and your named person has given the nod do we hand it over. Anything not yet quite right is recorded in the file with a date.
And if it does not work? Then there is the way back, and we agree it beforehand. As long as nobody has worked on the new server, we simply switch the old one back on. But once new orders have been entered, that would be a step back with data loss; the new data would first have to be transferred to the old server. That is why the old server stays in place for an agreed period after the move, with nobody writing to it, and is given a date on which it really goes.
09 / BACKUP AND RESTORING
Last week's file: can you get it back?
There are two very different emergencies. In the first, someone has overwritten a file and needs the state from the day before yesterday. In the second, the whole server is dead. The backup has to work for both, and for both we agree with you what you can tolerate: how far back do you need to be able to go, and how long can the business stand still?
That the backup reports “successful” every night is a start. Whether you can actually work from it is shown only by the test: we restore a version and check whether the file opens, whether the permissions are right and whether the database of the stock and order system starts (for the database together with the manufacturer, if necessary). We write the result down, with date and duration.
Databases are a special case here because they write constantly. A copy made in the middle of a write operation may, in the worst case, be unusable. That is why we back them up in the way the software manufacturer provides.
How we build backups (three copies, one of them off the premises) is described under Backup and disaster recovery. What a restore test shows and what it does not is explained in the guide to restore testing.
10 / UPDATES AND MAINTENANCE
Updates when they get in nobody's way.
A server needs care even when it runs quietly. Windows updates are pending, the disks are getting fuller, the backup did not complete last night. If we look after your server, it reports such things to us, and we deal with them so that they do not turn into a standstill. We do most of it by remote access from Rodewisch.
The procedure is always the same. We check what is due and how urgent it is. For anything that needs a restart, we agree a window with you in which it disturbs least, and we clarify beforehand how we would get back if an update causes trouble. Then we install it, restart and check that sign-in, drives and industry software are back. Finally it goes into the system file.
For us, an update only counts as done when it is installed and the server has restarted. “Downloaded” or “approved” does not count. And if the same warning keeps coming back (the disk is nearly full again), we do not just dismiss it every time but talk to you about more storage.
You can reach us Mon–Fri 8:00–17:00. Which response time applies and what ongoing support includes is set out in the support contract; the details are on the IT support page. If the industry software itself acts up, we bring in its manufacturer.
11 / FILE AND QUOTE
A server with a file.
Every server we set up or take over comes with the system file: what jobs it has, how it is set up, who may access what and how the backup runs. It stays with you, together with all the access credentials. So you could change service provider at any time without starting from zero.
It starts with a conversation, by phone or at your premises. That costs nothing. After that you receive a written quote in which survey, hardware, licences, set-up, migration and later support appear as separate items. So you can see what the device costs and what the work costs. We charge support by time spent or as a monthly flat rate; the rates are in the price list. If we order hardware for you, we agree a deposit for it.
We are based in Rodewisch and work throughout the Vogtland. We come to you for the survey, the installation and the migration; ongoing maintenance is mostly done by remote access.
Specification
DATA SHEET AT-02-S · SERVER IN OPERATION
- Jobs
- SHARED FOLDERS · INDUSTRY SOFTWARE · SIGN-IN AND PERMISSIONS
- System
- WINDOWS SERVER · ACTIVE DIRECTORY · DNS
- Hardware
- SIZED FROM MEASUREMENTS · TOWER OR RACK · RAID · UPS
- Migration
- OUTSIDE BUSY PERIODS · TESTED BEFOREHAND · WITH A WAY BACK
- Backup
- OLDER VERSIONS · RESTORE TESTED
- Support
- UPDATES · MONITORING · MAINTENANCE BY REMOTE ACCESS
- Documentation
- SYSTEM DOCUMENTATION · ACCESS STAYS WITH YOU
- Quote (Angebot)
- IN WRITING BEFORE WE START · INITIAL CONSULTATION FREE
- Available
- MON–FRI 8:00–17:00
Questions & answers
Common questions, answered briefly.
Do we still need a server of our own?
If your industry software needs a database on your premises, or everyone is meant to sign in at one central point: yes. If you only store files and already work with Microsoft 365, you can sometimes manage without. We look at what runs in your business, and we will tell you if you do not need one.
What happens if the server does not start on Monday morning?
Call us: +49 3744 4399760, Mon–Fri 8:00–17:00. On the phone we narrow down whether the cause is the server, the power or the network, and we come out if the hardware is defective. How quickly you are working again, though, is decided beforehand: by a backup whose restore has been tested, and by a file that records how the server was set up. For customers with a support contract, the response time promised there applies.
Is one server enough, or do we need two?
One is often enough, provided the backup has been tested and you can live with the time it takes to get going again (the test shows how long that is). A second one pays off if a single morning without your stock and order software costs more than the second machine. We work that out with you. Incidentally, one device can run several separate servers as virtual machines.
Our server is old but still running. Do we have to replace it?
Not because it is old. We check three things: does the Windows on it still receive updates, are spare parts still available for the hardware, and will your software soon need something newer? If all is well, it can keep running. If not, it is better to plan the replacement calmly now than under time pressure after a failure.
We have RAID. Does that mean our data is backed up?
No. RAID covers a failed disk and nothing more. A deleted or overwritten file is gone from all disks at the same time. For that you need a backup with older versions, ideally with a copy off the premises.
What happens to the server in a power cut?
Without a UPS it shuts off abruptly, and whatever was being written at that moment may be damaged. With a UPS (a battery in front of the server) it keeps running and, in a longer outage, shuts itself down cleanly. How long the battery lasts depends on the connected load and the age of the batteries. We size the UPS so that there is enough time for an orderly shutdown, and we test exactly that once.
An employee is leaving. What happens to their account?
You let us know, and we lock the account at the agreed time. It is only deleted later: first we clarify who takes over their files and whether a service is still running under their name. Locking can be done by remote access, so nobody has to come out.
Do you also look after our industry software?
The server underneath it, yes: set-up to the manufacturer's specification, updates, backup of the database, migration. The manufacturer is responsible for the program itself. If it is unclear which side a fault lies on, we coordinate with them directly so that you do not have to act as go-between.
Do we have to close the business for the server migration?
No. We build and test the new server while the old one keeps running. For the actual switchover we need a window in which nobody changes any data; we place it outside your busy periods. Afterwards your staff sign in as usual, only on the new server.
Our previous support person cannot be reached any more and nobody knows the passwords. Can you take over?
Yes, that happens more often than you might think. We gain access in an orderly way, record what runs on the server and create a system file. From then on the passwords exist again, and they are held by you.
What does a new server cost, including set-up?
That depends on the hardware, the licences and how much has to be moved; we can only give an honest figure after the survey. Before we start you receive a written quote (Angebot) with individual items, with ongoing support charged either by time spent or as a monthly flat rate. The initial consultation is free. For hardware orders we agree a deposit.
Where does your server support end?
At Linux servers and Proxmox. We work with Windows Server, virtualised with Hyper-V or VMware, and we do not take on anything we cannot do in depth. Questions about the industry software itself are answered by its manufacturer. And a server that never fails does not exist: the aim is that an outage stays short and that you know beforehand what to expect.
Contact
What is wrong with your server?
Tell us what you have and what bothers you. For business customers the initial consultation is free, by phone or at your premises. After that you receive a written quote.