1. Why conversion is not a simple copy and paste
A virtual machine consists of a disk, a simulated hardware configuration and an operating system that expects to find certain devices. When you change hypervisor, the disk format changes, as do the controllers through which the disk is presented, the network drivers and the components that connect the operating system to its environment. Even the firmware, the start-up software that initialises the machine, may differ between environments. For this reason, copying the disk is not enough on its own.
A converted machine may start up and behave differently from the original. Performance changes, network interfaces change name, addresses may be reassigned and disks are sometimes detected in a different order. Some software licences are tied to the machine's hardware identity. Each of these elements must be checked, because each can make a service fail even when the system appears to be working.
2. Prerequisites before you touch the machine
Before any conversion you must know exactly what you are moving. A disk snapshot without documentation produces a machine that nobody knows how to manage. Collect the operating system version, the list of installed applications, the active services and the network connections required. Also check that the operating system is still supported by the target environment and by the software vendor, because conversion does not resolve an expired support status.
- Full backup, verified with a restore test on a separate copy.
- Administrative access to the operating system, with documented service credentials.
- Target environment sized for the processors, memory, storage and network of the original.
- Maintenance window agreed with the managers of the application.
- Written rollback plan, with the maximum time within which to decide to revert.
3. The most common methods and when to use them
There are several ways to move a machine to another hypervisor. The first is to convert the virtual disk with dedicated tools, keeping the machine's configuration and adapting it to the new environment. The second is to export the machine in a standard exchange format and import it on the other side. The third is to rebuild the machine from scratch, installing a clean operating system and migrating data and applications. The choice depends on the complexity of the machine, the downtime allowed and the amount of undocumented configuration it carries.
The table summarises the trade-offs. Direct conversion is the quickest when the system is simple, but drivers must be fixed after the move. Import from a standard format is easier to repeat and document, but depends on support for the format at both ends. Rebuilding is the cleanest, but also the longest and the most exposed to configuration differences. When the machine has years of manual changes, rebuilding often makes more sense than converting.
| Method | Main advantage | Limit to consider |
|---|---|---|
| Disk conversion | Keeps the machine's configuration and data | Drivers and integration tools must be fixed afterwards |
| Import from standard format | Repeatable and documentable procedure | Depends on format support at both ends |
| Clean installation and data migration | Clean environment, free of legacy residue | More work and greater risk of configuration differences |
4. Operational steps
The following sequence applies to direct disk conversion, which is the most frequent case. Adapt it to the method you choose, but do not skip the isolated start-up phase, because it is what protects production from errors. Each step has an expected result that must be recorded, so that whoever intervenes later can understand the state of the machine. The person carrying out the procedure should keep the list at hand and tick each step before moving to the next.
Never connect the converted machine to the live network before its first controlled start. A machine with the same name and address as the original can cause conflicts, authentication errors or writes to data it should not touch. Use an isolated network for the first checks, and connect the machine to the production network only once the result has been approved.
- Stop the application services in an orderly way and create a snapshot or a consistent copy of the disk.
- Convert the disk into the format required by the new hypervisor, without altering its contents.
- Create the new machine with firmware, processors and memory consistent with the original.
- Attach the converted disk and check the order of disks and controllers before the first start.
- Start the machine on an isolated network, with no links to production systems.
- Install the integration tools provided by the new hypervisor and remove those of the old one.
- Fix the network configuration, interface names, mount paths and licences.
- Restart the services and connect the machine to the live network only after the checks.
5. Common mistakes to avoid
The first mistake is considering the conversion successful because the machine boots, when a boot only shows that the operating system can find a disk. The second is forgetting that addresses, interface names and mount paths may change, with services connecting to an old address. The third is leaving drivers and tools from the old hypervisor, which cause errors and instability. The fourth is neglecting licences tied to hardware identity, which must be checked with the vendor beforehand, not afterwards. The fifth is converting machines that should be rebuilt, because they carry years of undocumented configuration.
6. How to verify that the conversion succeeded
Verification must be carried out on several levels. Check that the operating system starts fully, that services are running and that logs show no recurring errors. Compare the contents with the original, using file counts and checksum comparisons on samples or critical paths. The most useful test remains the functional one: a complete work process run by the people who use it every day, such as recording an order or producing a printout. Document the result, because it will be the basis for deciding when to decommission the original machine.
7. What to do after the conversion
Keep the original machine preserved for an agreed period, so that you can revert if problems emerge that the tests did not reveal. Monitor the new machine in the first few days, paying attention to performance, errors and response times. Update the inventory and check that backups resume, because backup policies often depend on the environment identifier or an installed agent, which may no longer be present after conversion.
8. How to start
This week, choose a non-critical test machine running a simple application. Make an inventory of its components, run a backup with a restore test, then convert the copy in an isolated environment, noting every difference you come across. Ask the provider of the target environment which disk formats it supports, which integration tools are available and what procedure it prescribes for licences, with answers in writing.
Frequently Asked Questions
Can I convert a virtual machine without stopping it?
It depends on the method and the application. A copy taken from a consistent snapshot can reduce downtime, but for data that changes constantly you need a procedure that guarantees consistency, otherwise the disk starts with incomplete data. The downtime you can tolerate should be agreed with the application managers before you choose the method.
Does conversion keep the data?
Usually yes, if the process is carried out correctly and the disk has been copied consistently. The risk, however, is not limited to content: it also concerns configuration changes. That is why a comparison is needed after conversion, not merely a check that the machine boots.
Do I need to change the machine's operating system?
Not necessarily, but you must verify that the version is supported by the new hypervisor and by the software vendor you use. Some very old versions are no longer supported everywhere. In that case conversion is not the right solution: consider rebuilding on a supported version, restoring data and applications.
Do I need a specialist technician?
For simple machines with clear documentation, an in-house technician with virtualisation experience may be enough. For machines running critical applications, complex licences or sensitive data, external support reduces the risk. Ask the provider for examples of similar conversions and for the verification plan it proposes before entrusting the machine to it.
How long does a conversion take?
It depends on the size of the disks, the speed of the network and the method chosen. There is no reliable general estimate. Measure the time on a test machine with data similar to your real data, and use that figure to plan the cut-over window, adding a margin for checks and any rollback.