No description
Find a file
2024-07-02 20:28:31 -07:00
defaults Startup nebula ansible role 2024-04-18 18:30:15 -07:00
handlers Startup nebula ansible role 2024-04-18 18:30:15 -07:00
meta Startup nebula ansible role 2024-04-18 18:30:15 -07:00
tasks Startup nebula ansible role 2024-04-18 18:30:15 -07:00
templates add new option for v1.9.X series 2024-07-02 20:28:31 -07:00
tests Startup nebula ansible role 2024-04-18 18:30:15 -07:00
vars Startup nebula ansible role 2024-04-18 18:30:15 -07:00
.ansible-lint Startup nebula ansible role 2024-04-18 18:30:15 -07:00
.yamllint.yml Startup nebula ansible role 2024-04-18 18:30:15 -07:00
README.md Startup nebula ansible role 2024-04-18 18:30:15 -07:00

ansible-role-nebula

This role helps setup Nebula on applicable nodes in the RESF. Most settings for this role are specifically for the RESF and its projects. However, it is perfectly possible to use this on your own. Note that this relies specifically on rocky-release-core being installable, which means Rocky Linux will work without issues. Fedora Linux will work as nebula is available in their base repositories. Other distributions may not work.

If there are issues with this role for your use case, please file an issue or a PR if you would like to enhance this role.

Requirements

Requirements are as follows:

  • Enterprise Linux 9+ or Fedora Linux
  • Ansible collections: community.general
  • ansible-core >= 2.14

Role Variables

A description of the settable variables for this role should go here, including any variables that are in defaults/main.yml, vars/main.yml, and any variables that can/should be set via parameters to the role. Any variables that are read from other roles and/or the global scope (ie. hostvars, group vars, etc.) should be mentioned here as well.

Dependencies

  • community.general
  • ansible.posix (only for tests)

Variables

This is not an all inclusive list. For additional variables, check defaults/main.yml.

Variable Default Value Required Description
nebula_am_lighthouse false Conditional Sets this node as a lighthouse.
nebula_lighthouse_internal 60 No How often (in seconds) should a node report to a lighthouse.
nebula_routable_ip No The public routable IP that nebula needs to know about. If not set, it will be determined automatically.
nebula_ip Yes IP required by nebula in the form of X.X.X.X/X (for example, 10.100.0.44/24).
nebula_ca_host Yes The hostname of the host which should be used as a CA. Exactly one (1) MUST be a CA. Required if nebula_am_lighthouse is true.
nebula_is_ca false Conditional If the host is the CA or not. If nebula_ca_host is not defined, exactly one play host will need this set to true. Required if nebula_am_lighthouse is true.
nebula_is_member true Yes This node a member of the mesh.
nebula_ca_name RESF Nebula CA Yes Sets the name of the CA.
nebula_ca_life 175200h No Sets the life of the CA certificate.
nebula_ca_wait_timeout_secs 300 No Timeout in seconds for members to wait until the CA is ready to issue certificates.
nebula_nodename {{ ansible_facts.hostname }} No Name of this nebula member. This is determined by the hostname by default. Otherwise, it can be set.
nebula_groups [] Conditional List of groups that a node is assigned to. This added to the issued certificate for the node.
nebula_listen_host 0.0.0.0 Conditional The IP of the interface nebula will need to bind to. Default is all IPv4 interfaces. Use [::] if you want to enable IPv6.
nebula_listen_port 4242 Conditional The port to bind to. Default is 4242, just like the documentation.
nebula_listen_batch No Max number of packets to pull from the kernel on each syscall.
nebula_listen_read_buffer No Read socket buffers for the UDP side. Values will be doubled in the kernel. Default is net.core.rmem.default on the system.
nebula_listen_write_buffer No Read socket buffers for the UDP side. Values will be doubled in the kernel. Default is net.core.wmem.default on the system.
nebula_listen_send_recv_error No Nebula will reply to packets it has no tunnel for with a recv_error packet. This helps speed up reconnection in cases of nebula not shutting down cleanly. The caveat is that this can be used to abuse checking for nebula running on some host.
nebula_punchy_punch true Conditional Used for NAT situations. Most cases NAT exists, so this is set to true. Enabling this causes the node to send small packets at an interval.
nebula_punchy_respond No Set this to true if the node is unable to receive handshakes and will attempt to initiate one (in the case where hole punching fails in one direction). Useful if a host is behind a difficult NAT (like symmetric NAT).
nebula_punchy_respond_delay No Set this to the number of seconds to delay before attempting a punch. Only valid is nebula_punchy_respond is true.
nebula_punchy_delay No Set this to the number of seconds to delay/slow down punch responses. This is helpful if NAT is unable to handle certain race conditions. Only valid if nebula_punchy_respond istrue.
nebula_cipher aes No Unless you know what you're doing, avoid touching this setting. Refer to the nebula documentation.
nebula_tun_disabled false Conditional Set to true if you do not want the tunnel up. Most people want a tunnel.
nebula_tun_dev rneb01 No Set the tunnel device name.
nebula_tun_drop_local_broadcast false No Toggles forwarding of local broadcast packets. This depends on the CIDR in the certificate for the node.
nebula_tun_drop_multicast false No Toggles forwarding of multicast packets.
nebula_tun_tx_queue 500 No Transmit queue length. Raise this number if there are a lot of transmit drops.
nebula_tun_mtu 1300 No Default MTU for every packet. Safest setting is 1300 for internet routed packets.
nebula_tun_use_system_route_table false No Exactly as it says, set to true if you want to manage unsafe routes directly on the system route table with gateway routes instead of nebula.
nebula_routes [] No List of dictionaries. Use this to create route based MTU overrides. If you have a known path that can support a larger MTU, you can set it this way.
nebula_unsafe_routes [] No List of dictionaries. Allows you to route traffic over nebula to non-nebula nodes. This should be avoided unless you have hosts that cannot run nebula. See nebula documentation.
nebula_logging_level info No Sets the log level.
nebula_logging_format text No Formatting of the logs. Can be text or json.
nebula_logging_disable_timestamp false No Disables timestamp logging. If the output is redirected to some logging system, set to true.
nebula_logging_timestamp_format No Sets the timestamp format. Default is RFC3339 unless format is text and is attached to a TTY.
nebula_firewall_conntrack_tcp_timeout 12m No Sets the connection tracking TCP timeout.
nebula_firewall_conntrack_udp_timeout 3m No Sets the connection tracking UDP timeout.
nebula_firewall_conntrack_default_timeout 10m No Sets the default connection tracking timeout.
nebula_firewall_inbound_rules any No List of dictionaries. Sets the appropriate inbound rules for this node.
nebula_firewall_outbound_rules any No List of dictionaries. Sets the appropriate outbound rules for this node.
nebula_pki_disconnect_invalid true No Forcefully disconnets a client if the certificate is expired or invalid.
nebula_pki_block_list [] No List of certificate fingerprints that should be blocked even if it's valid.
nebula_cert_public_key No Nebula node public key to use. If defined, no public key is generated on the CA. This will be signed and used. Requires nebula_cert_private_key to be set.
nebula_cert_private_key No Nebula node privatekey to use. If defined, no public key is generated on the CA. This will be used. Requires nebula_cert_public_key to be set.
nebula_preferred_ranges No Sets the priority order for underlay IP addresses. See the documentation.
nebula_routines No Number of thread pairs to run that consume from the tun and UDP queues. The default is 1, which means there's one tun and one UDP queue reader. The maximum recommended setting is half the available CPU cores.

Example Playbook

License

...

Author Information

Louis Abel label@rockylinux.org