id: PG-10
category: networking
title: ACTIVE, but no IP address?
situation: >-
  A new instance is `ACTIVE`, no fault, yet unreachable. Instances created earlier work fine. Its DHCP
  request got no answer.

diagram:
  type: chain
  caption: Nova did its job. The failure comes after
  nodes:
    - label: Nova
      sub: "Instance `ACTIVE`, power state running"
      icon: cloud
      tone: ok
    - label: Guest boots
      sub: Asks for a lease, gets no answer
      icon: monitor
      tone: warn
      badge: clock
    - label: dnsmasq
      sub: Does not know the new port
      icon: network
      tone: mute
    - label: DHCP agent
      sub: Not alive, so the port was never added
      icon: server
      tone: stop
  links: [ok, broken, broken back]
  note: >-
    Running instances keep their lease (`dhcp_lease_duration`, 86400 s by default). On the Nova side,
    only the console log shows the failure.

checks:
  - label: No fault in Nova
    hint: openstack server show <id> -c status -c fault -c addresses
  - label: "Console: DHCP asked, no answer"
    hint: openstack console log show <id> | grep -iE 'dhcp|lease|bound'
  - label: Is any DHCP agent alive?
    hint: openstack network agent list --agent-type dhcp
  - label: Agents serving that network
    hint: openstack network agent list --network <network>

caution:
  - level: know
    text: >-
      A `qdhcp-` namespace with dnsmasq still running is not proof of health. Check that the agent itself is
      alive.
  - level: state-changing
    text: >-
      Instances created during the outage may need a reboot to get a lease. Tell their owner first.

lesson: ACTIVE means Nova did its job. It says nothing about the network of the guest.

runbook:
  path: networking/new-instance-gets-no-ip-dhcp-agents-down.md
  status: lab-verified
  release: "2026.1"

scope: >-
  Applies to ML2/OVS with DHCP agents. Lab-verified on OpenStack 2026.1 with Kolla-Ansible for the case where
  every DHCP agent is down; a single network without a live agent and a missing dnsmasq were not reproduced.
  OVN has no DHCP agent: this memo does not apply (not tested). A subnet with DHCP disabled gives the same
  symptom. Starting the agents again
  can move networks to another DHCP agent (observed on a lab, without traffic loss).

linkedin:
  title: "ACTIVE, no fault, no IP address: Nova did its job"
  caption: |
    A new instance is ACTIVE. No fault. Nothing in the Nova logs. And nobody can reach it.

    Instances created yesterday are fine, which makes it look like a problem with this one VM. It is not.

    With ML2/OVS, a new port must be added to dnsmasq by a DHCP agent. If no agent is alive, the guest asks for a lease and never gets an answer. Running instances keep theirs (24 hours by default), so the outage is invisible until someone creates an instance.

    Where to look:
    → the console log of the guest: DHCP asked, never answered
    → openstack network agent list --agent-type dhcp

    ACTIVE means Nova did its job. It says nothing about the network of the guest.
  hashtags: [OpenStack, Neutron, Networking, CloudOperations, SRE, OzarLand]
