Skip to content

Troubleshooting

Symptom → fix, in the order you'll likely meet them. When in doubt, the task log (bottom panel — double-click any task for its full output) is always the first place to look.

Can't reach the web UI

Symptom: https://192.168.1.50:8006 doesn't load.

  • Is that the right IP? The installer prints it at the end; check your router's client list if you're unsure.
  • Is it https, not http? Port 8006 serves HTTPS only — plain http:// fails silently-ish.
  • Did you accept the certificate warning? Self-signed on a fresh install is expected.
  • Can you ping the host? If not, it's a network problem, not a Proxmox problem: cable, switch port, IP conflict with another device squatting on the static address you chose.

apt update errors about repositories

Symptom: errors mentioning enterprise.proxmox.com, possibly 401s.

You still have the enterprise repository enabled without a subscription key. See Installing Proxmox — switch to the no-subscription repository. This is the single most common fresh-install issue.

VM won't start

Symptom: clicking Start produces a failed task.

  • Read the task log (double-click it). The error is usually specific: "no such logical volume," "bridge 'vmbr0' does not exist," "not enough memory."
  • "No bootable disk": the ISO isn't attached or the install never completed — check the VM's Hardware tab for the CD/DVD drive.
  • Check host resources: node → Summary. A host with no free RAM can't start guests.

Guest has no network

Symptom: VM runs, but no IP, no internet.

  • Is the guest's network device attached to vmbr0? (VM → Hardware → network device.)
  • Inside the guest, is DHCP actually running? A static-IP install with the wrong gateway looks exactly like "no network."
  • On the host, does vmbr0 itself have the physical NIC attached? (Node → Network.) A bridge with no uplink is a switch plugged into nothing.

Storage full

Symptom: new VMs fail to create, backups fail, snapshots error out.

  • Node → Storage shows usage per storage. Backups accumulating on local is the usual culprit — old backup files don't delete themselves unless retention is set (Datacenter → Backup → your job → Retention).
  • Snapshot hoarding also eats space: qm listsnapshot and delete the ones whose experiment is over.

A VM is "locked"

Symptom: operations fail with a lock error (e.g. after a crashed backup or migration).

The task log shows what locked it. If the underlying task is genuinely dead (not just slow — check!), the UI offers an unlock option on the VM, and qm unlock <vmid> exists on the command line (verified against the official qm man page). Unlock only when you're sure nothing is actually running — unlocking a live operation corrupts things.

Container won't start: unprivileged and mounts

Symptom: a container fails immediately on start.

Unprivileged containers can't do some things privileged ones can (certain mounts, device passthrough). If you followed a tutorial that assumed privileged mode, that's your mismatch. Prefer fixing the approach over flipping the container to privileged.

The meta-fix

Proxmox logs everything. Node → System Log, the per-task logs, and journalctl on the host shell answer most questions. When asking for help (forum, IRC), include: VE version (pveversion), what you clicked/typed, and the full task log text. That's the difference between help in minutes and help never.