I recently ran into a frustrating issue with my Proxmox backups. Across my three Proxmox hosts, one was consistently failing, and the error message pointed to a peculiar problem with a specific VM (VM 103).
The backup logs showed this:
103: 2025-03-28 06:51:46 INFO: Starting Backup of VM 103 (qemu)
103: 2025-03-28 06:51:46 ERROR: disk image '/mnt/pve/TN_SSD1/images/103/vm-103-fleece-0.qcow2' already exists
103: 2025-03-28 06:51:46 ERROR: Backup of VM 103 failed - disk image '/mnt/pve/TN_SSD1/images/103/vm-103-fleece-0.qcow2' already exists
Essentially, the backup process was failing because it detected an existing disk image with the same name. A quick check of the VM’s directory confirmed this - I had two disk images:
vm-103-fleece-0.qcow2vm-103-disk-0.qcow2
The odd thing? The vm-103-fleece-0.qcow2 file only appeared for this VM and wasn’t reflected in the VM’s hardware configuration within the Proxmox web interface. I could only see one disk listed!
I tried everything I could think of: deleting old backups, specifically targeting the failing backup, even removing all backups for that VM. Nothing worked – the error persisted. Attempts to delete the phantom disk directly resulted in a “disk in use” message.
The Solution: qm rescan to the Rescue!
Finally, I stumbled upon the fix: the qm rescan command. Running qm rescan --vmid 103 revealed the problem. Proxmox identified the missing disk and flagged it as "unused."
rescan volumes...
VM 103 add unrefreferenced volume 'TN_SSD1:103/vm-103-fleece-0.qcow2' as 'unused0' to config
With the disk now visible in the Proxmox configuration, I could access the VM’s hardware settings and finally remove the unattached disk.
After deleting the phantom disk, the scheduled backup job completed successfully!
If you encounter a similar issue with Proxmox backups and a “disk image already exists” error, don't just focus on deleting backups. Run qm rescan --vmid [VMID] to ensure Proxmox is aware of all disk images associated with the VM. It might just save you a lot of headaches!
