Skip to main content

Troubleshooting — Virtual machines

The Next button stays greyed out in the wizard​

Cause: the VM would exceed a project quota (CPU, Memory or Storage). The wizard banner shows Quota exceeded and the breakdown per resource.

Solution:

  1. Choose a smaller instance type or reduce the size of the disks.
  2. Free up resources: delete unused VMs or disks (detached disks count against the storage quota).
  3. Have the project quotas increased.

If the console says the project quotas are unavailable, creation stays blocked until they can be read: try again later or contact support.

Error at creation: name already in use or disk size rejected​

Cause and solution:

MessageSolution
An instance with this name already exists.Choose another name.
Min. 20 GB required, Min. 28 GB required, Min. 40 GB required or This image needs a disk of at least … GBIncrease Size (GB): 20 GB minimum; for the system disk, 28 GB with Windows Server and 40 GB with Oracle Linux.
A system image is required for a new diskSelect a card under Operating System.
Invalid format. Expected: <algorithm> <base64-key> [comment]Paste the complete public key (.pub file), on a single line.

Saving changes to a VM is refused​

Cause: a field on the Edit page is invalid, for example a new disk under 20 GB or an Existing volume with no disk selected.

Solution: the console shows Incomplete form with the message of each invalid field. Fix those fields, then click Save again. Disks already attached to the VM are shown read-only and are not held to the rules of a new disk (name, minimum size).

The system disk cannot be replaced, and New / Existing does not appear on an attached disk​

Cause: on the Edit page, the disks the VM already has are read-only. The System Disk (Boot) has no action ("The system disk of an existing VM cannot be replaced: the VM boots from it.") and a data disk only offers Detach disk. The New / Existing choice only exists on disks added with Add a disk.

Solution:

  1. To replace a data disk, detach it, add a disk with Add a disk (New or Existing), click Save, then Restart.
  2. To change the system disk, delete the VM (its disks are kept), then recreate it choosing Existing on the System Disk (Boot) (see the FAQ).

The VM stays in Error or Failed status​

Cause: the VM could not be scheduled or started (resources unavailable, disk in error, GPU unavailable…). When the platform returns a reason, it is shown when you hover over the status badge.

Solution:

  1. Open the detail page and check the Storage & Disks section: the disks must be present.
  2. If the VM has GPUs, see GPU unavailable at startup.
  3. Try Stop then Start from the Actions section.
  4. If the status persists, contact support, giving the VM name and its identifier (shown under the title of the detail page, with a copy button).

SSH timeout​

Cause: no public IP, port 22 not allowed, or SSH service not yet started in the VM.

Solution:

  1. On the detail page, Network & Security section: Public IP must be Active and port 22 must appear under Firewall & Ports.
  2. Otherwise, click Edit, enable Public IPv4 Address, check SSH (22) in Allowed Ports, then Save.
  3. Right after creation, wait one to two minutes for the OS to finish booting.
  4. Test in verbose mode: 
    ssh -v ubuntu@<public-ip>

Permission denied (publickey)​

Cause: wrong user, wrong key, or key added after the first boot without reloading the user-data.

Solution:

  1. Use the user shown in the SSH Connection block (or under System Image > User).
  2. Check that the public key matching your private key appears in Advanced Configuration > SSH Keys.
  3. If you just added the key via Edit, choose Reload user-data in the SSH keys changed dialog, or run Reload UserData from the Actions section, then Restart: the key is only installed at restart. This restart also regenerates the SSH host keys: if your client reports a changed key, remove the old entry with ssh-keygen -R <public-ip>.

The added disk does not appear in the VM​

Cause: the VM has not restarted yet after the addition, or the disk is not formatted.

Solution:

  1. After Save, the console shows Restart required: the disk is only added on the next restart. Click Restart in the Actions section, then wait for the VM to return to the Running state.
  2. Check the disk in the Storage & Disks section of the detail page.
  3. In the VM, list the devices: a new disk appears with no partition and no mount point. 
    lsblk
  4. Format and mount it: see Attach an extra disk.

GPU unavailable at startup​

Cause: a GPU is released when the VM is stopped and may be assigned to another workload in the meantime.

Solution: at startup, if the GPU is no longer available, the console opens the Select an alternative GPU dialog. Choose a model in Available GPU, then click Update and Start. If the dialog says No GPUs are currently available., try again later or contact support. See GPU troubleshooting.

.local DNS does not work in the VM​

Cause: systemd-resolved treats .local domains as mDNS.

Solution: see Resolve .local DNS in VMs.

The VM does not respond at all (neither SSH nor RDP)​

Cause: OS hung, network misconfigured inside the VM, internal firewall too restrictive.

Solution:

  1. Run Restart from the Actions section of the detail page.
  2. If a recent cloud-init change is the cause, fix the script in Edit > Advanced Configuration, then run Reload UserData and Restart.
  3. Serial console or VNC access is not offered in the console; contact support for low-level diagnostics.