id: PG-11
category: networking
title: Reachable inside, not outside?
situation: >-
  The instance has its address and its metadata, and a neighbour in the tenant network reaches it. From
  outside the cloud: a timeout, not a refusal.

diagram:
  type: chain
  caption: Four checks, in this order
  nodes:
    - step: "1"
      label: Security group
      sub: An ingress rule for this protocol, port and source?
      icon: shield
      tone: ink
    - step: "2"
      label: Floating IP
      sub: On this port, and this fixed address?
      icon: map-pin
      tone: ink
    - step: "3"
      label: Router
      sub: Gateway, interface, and live on its node?
      icon: route
      tone: ink
    - step: "4"
      label: Guest
      sub: Host firewall, service listening?
      icon: monitor
      tone: ink
  note: >-
    Most cases end at 1 or 2. If every floating IP stops answering at once, the cause is shared: the gateway
    node or the external network.

checks:
  - label: Port and its security groups
    hint: openstack port show <port> -c security_group_ids
  - label: Ingress rules of each group
    hint: openstack security group rule list --ingress <group>
  - label: A floating IP on this port?
    hint: openstack floating ip list --port <port>
  - label: Router external gateway
    hint: openstack router show <router> -c external_gateway_info

caution:
  - level: know
    text: >-
      A rule whose remote is a security group admits only members of that group. Neighbours get in, the
      outside does not.
  - level: state-changing
    text: >-
      A new rule opens every port that uses the group. Use the narrowest source prefix that works.

lesson: Each check rules out one cause. Go through them in order.

runbook:
  path: networking/instance-unreachable-from-outside.md
  status: lab-verified
  release: "2026.1"
  also:
    - networking/router-stuck-on-dead-l3-agent.md

scope: >-
  Written from the Neutron 2026.1 documentation and lab-verified on ML2/OVS with the L3 agent and a legacy
  router, for the security group and floating IP causes. The OVN steps, the router missing on its node, the
  guest and the external network causes were not run on a lab. DVR, BGP-announced floating IPs and IPv6
  without NAT, and provider networks attached directly to instances are not covered.

linkedin:
  title: "Reachable from inside, not from outside: four checks, in order"
  caption: |
    The instance got its address and its metadata. A neighbour in the same network reaches it. From outside: timeout.

    This is rarely a mystery. It is four things to check, in order:
    1. Security group: is there an ingress rule for this protocol, port and source?
    2. Floating IP: is it on this port, and on this fixed address?
    3. Router: an interface and a gateway, and is it really live on its node?
    4. The guest: its own firewall, a service actually listening?

    Most cases end at 1 or 2.

    One classic trap: an ingress rule whose remote is a security group, not a prefix. It admits members of that group only. That is exactly why the neighbour gets in and the outside does not.
  hashtags: [OpenStack, Neutron, Networking, SecurityGroups, CloudOperations, OzarLand]
