A virtual machine is deleted from the Proxmox interface in two clicks, but the action is irreversible: its disks disappear together with the machine unless you detach them first. Backups and snapshots kept on other storage are not touched, and that is a frequent source of confusion.
What to check before deleting a virtual machine
Before clicking delete, check everything that might reference the machine or live separately from it — otherwise part of the data and settings stays behind for no visible reason.
- Snapshots — they are stored with the disk and disappear together with it
- Backups — they live separately and do not keep the machine from being removed
- Disks attached from other storage, separate from the main one
- Datacenter and node firewall rules tied to the machine
- DNS and monitoring records pointing to the machine address
Deleting a virtual machine step by step
The machine is stopped first and only then removed — this is deliberately not a single button, so a running service cannot be torn down by accident.
- Stop the machine (not just shut down the guest, but stop it in Proxmox itself)
- Choose the delete option and decide what happens to the disks
- Confirm the action — after this it cannot be undone
qm stop VMID # Stop the machine
qm destroy VMID --purge # Delete the machine together with its disks and recordsThe disk checkbox decides everything: leave it checked and the disks are wiped together with the machine; clear it and the machine disappears while the disks remain as orphaned volumes you have to remove or reuse manually.
What remains after deletion and where to look
Some objects are not tied to the machine directly and survive its deletion — look for them separately if you need to free space or clean things up.
| Object | Where it lives | Deleted automatically |
|---|---|---|
| Backups in Proxmox Backup Server | Backup storage | No |
| Snapshots made on the node's local storage | Local disk of the node | No |
| A disk attached from separate storage | NFS, Ceph, or an external volume | No |
| A firewall rule and a DNS record | Datacenter configuration and the DNS zone | No |
Why removing a node from a cluster is riskier than deleting a machine
A Proxmox cluster joins several physical nodes into one management system — for more on the platform itself see the overview what Proxmox VE is. A node in that system is not simply switched off: it holds part of the shared configuration and the quorum state.
A removed node cannot be brought back under the same name without extra steps — the system remembers the old record and treats it as conflicting with the new one. That is why removing a node is planned in advance, not done after something already broke.
Order of removing a node from a cluster
A node is removed in a specific sequence, and every step is run from another, still working node of the cluster — never from the node being removed.
- Migrate all guest machines and containers off the node being removed to the remaining ones
- Remove the node with a command run from any other cluster node
- Do not bring the removed node back into the same network under its old name
pvecm delnode NODENAME # Run from another node of the clusterIf a node is removed in an emergency, without migrating guests first, their machines on it become unreachable until manually restored — this is a case where the order of steps matters more than speed.
Dissolving the whole cluster
Dissolving an entire cluster is rarely needed — usually when moving back to standalone nodes or rebuilding the infrastructure from scratch, not for a one-off problem with a single node.
Unlike removing a single node, this returns every node to an independent setup: each node keeps its own configuration, and the shared quorum and the shared configuration database stop existing.
Frequently asked questions
Are disks deleted together with the machine?
Yes, if the disk-destroy checkbox is left checked during deletion, the disks are wiped together with the machine with no way to recover them. If the checkbox is cleared, the disks remain on storage as separate volumes that still need to be removed or reused manually afterward.
Can a deleted machine be brought back?
The machine itself cannot — deletion is irreversible and there is no undo in the interface. Data can only be restored from a backup, provided one was made in advance and stored separately from the deleted machine, for example in Proxmox Backup Server or on another storage location.
How do you remove a node from a cluster?
First the guest machines are migrated off that node to the remaining ones, then the node is removed from another, still working node of the cluster. The removed node is not brought back into the same network under its old name, since the system remembers the old record and treats it as a conflict.
What happens to guests on the node being removed?
If guests are not migrated beforehand, they become unreachable once the node is removed and do not start automatically on other nodes. Migration has to happen before the node leaves the cluster, not after — recovering from an emergency removal takes noticeably longer.
Conclusion
Deleting a machine and removing a node carry a different cost of error: the first wipes disks immediately, the second breaks configuration at the cluster level. Before either one, check snapshots, backups, and dependent records, and for simple isolated setups without a cluster consider LXC containers, where there is less chance of touching someone else's data by accident.