How to detach additional volumes from virtual machines

You can detach a secondary disk or a shared filesystem from a VM. You cannot detach a boot disk.

Steps

Get the disk’s device ID or the filesystem’s mount tag

Web console
CLI

In the sidebar, go to Storage → Disks or Storage → Shared filesystems.
Copy the device ID from the list of disks or the mount tag from the list of filesystems.

Get the VM’s specification:

nebius compute instance get <VM_ID>

Alternatively, list all VMs:

nebius compute instance list

In the output, you can find device IDs of disks and mount tags of filesystems in .spec.secondary_disks[].device_id and .spec.filesystems[].mount_tag, respectively:

spec:
  secondary_disks:
    - attach_mode: READ_WRITE
      existing_disk:
        id: computedisk-***
      device_id: newdisk
  filesystems:
    - attach_mode: READ_WRITE
      existing_filesystem:
        id: computefilesystem-***
      mount_tag: shared-fs

Make sure that the VM does not mount the volume when the VM restarts

Connect to the VM over SSH.
Switch to the root user:
sudo su -
If you want to detach a secondary disk, get its UUID:
blkid /dev/disk/by-id/virtio-disk-0-part1 -o export | grep "^UUID"
In this example, we assume that you have named the device disk-0; for disks, all device IDs are prefixed with virtio-.
Open /etc/fstab with your preferred text editor, for example, nano /etc/fstab, and delete the lines that refer to the volumes:
- The line for a disk should start with its UUID:
  UUID=<disk_UUID> /mnt/disk-0 ext4 defaults,nofail 0 2
- The line for a filesystem should start with the filesystem’s mount tag from the VM specification:
  filesystem-0 /mnt/fs virtiofs rw,nofail 0 0
  In this example, the filesystem has the mount tag filesystem-0 and is mounted at /mnt/fs.
Disconnect from the VM.

Remove the volume from the VM’s specification

Web console
CLI

In the sidebar, go to Compute → Virtual machines.
Open the page of the required VM.
Click Stop VM and then confirm it.
You can detach a disk from a running VM, but this can cause data loss or corruption.
After the VM’s status becomes Stopped, switch to the Disks or Filesystems tab.
Next to the disk or filesystem, click → Detach.
After the volume is detached, click Start VM and then confirm it.

Stop the VM:
nebius compute instance stop --id <VM_ID>
You can detach a disk from a running VM, but this can cause data loss or corruption.
Start editing the VM specification:
nebius compute instance edit <VM_ID>

In the text editor that opens, remove the disk or filesystem from the .spec.secondary_disks or .spec.filesystems list, respectively. If, after that, a list becomes empty, remove it entirely. For example:

  spec:
    boot_disk:
      attach_mode: READ_WRITE
      device_id: ""
      existing_disk:
        id: computedisk-***
-   secondary_disks:
-     - attach_mode: READ_WRITE
-       existing_disk:
-         id: computedisk-***
-       device_id: newdisk
    gpu_cluster:
  ...

After making the changes, save the file. The CLI updates the VM automatically.

Start the VM:

nebius compute instance start --id <VM_ID>

You can attach and mount different additional volumes to the VM later.

Virtual machines

Slurm and Soperator in Nebius AI Cloud

Managed Service for Kubernetes®

How to detach additional volumes from virtual machines

Steps

Get the disk’s device ID or the filesystem’s mount tag

Make sure that the VM does not mount the volume when the VM restarts

Remove the volume from the VM’s specification

​Steps

​Get the disk’s device ID or the filesystem’s mount tag

​Make sure that the VM does not mount the volume when the VM restarts

​Remove the volume from the VM’s specification

Steps

Get the disk’s device ID or the filesystem’s mount tag

Make sure that the VM does not mount the volume when the VM restarts

Remove the volume from the VM’s specification