kolla-ansible/tests/pre.yml
Radosław Piliszek bc053c09c1 Implement IPv6 support in the control plane
Introduce kolla_address filter.
Introduce put_address_in_context filter.

Add AF config to vars.

Address contexts:
- raw (default): <ADDR>
- memcache: inet6:[<ADDR>]
- url: [<ADDR>]

Other changes:

globals.yml - mention just IP in comment

prechecks/port_checks (api_intf) - kolla_address handles validation

3x interface conditional (swift configs: replication/storage)

2x interface variable definition with hostname
(haproxy listens; api intf)

1x interface variable definition with hostname with bifrost exclusion
(baremetal pre-install /etc/hosts; api intf)

neutron's ml2 'overlay_ip_version' set to 6 for IPv6 on tunnel network

basic multinode source CI job for IPv6

prechecks for rabbitmq and qdrouterd use proper NSS database now

MariaDB Galera Cluster WSREP SST mariabackup workaround
(socat and IPv6)

Ceph naming workaround in CI
TODO: probably needs documenting

RabbitMQ IPv6-only proto_dist

Ceph ms switch to IPv6 mode

Remove neutron-server ml2_type_vxlan/vxlan_group setting
as it is not used (let's avoid any confusion)
and could break setups without proper multicast routing
if it started working (also IPv4-only)

haproxy upgrade checks for slaves based on ipv6 addresses

TODO:

ovs-dpdk grabs ipv4 network address (w/ prefix len / submask)
not supported, invalid by default because neutron_external has no address
No idea whether ovs-dpdk works at all atm.

ml2 for xenapi
Xen is not supported too well.
This would require working with XenAPI facts.

rp_filter setting
This would require meddling with ip6tables (there is no sysctl param).
By default nothing is dropped.
Unlikely we really need it.

ironic dnsmasq is configured IPv4-only
dnsmasq needs DHCPv6 options and testing in vivo.

KNOWN ISSUES (beyond us):

One cannot use IPv6 address to reference the image for docker like we
currently do, see: https://github.com/moby/moby/issues/39033
(docker_registry; docker API 400 - invalid reference format)
workaround: use hostname/FQDN

RabbitMQ may fail to bind to IPv6 if hostname resolves also to IPv4.
This is due to old RabbitMQ versions available in images.
IPv4 is preferred by default and may fail in the IPv6-only scenario.
This should be no problem in real life as IPv6-only is indeed IPv6-only.
Also, when new RabbitMQ (3.7.16/3.8+) makes it into images, this will
no longer be relevant as we supply all the necessary config.
See: https://github.com/rabbitmq/rabbitmq-server/pull/1982

For reliable runs, at least Ansible 2.8 is required (2.8.5 confirmed
to work well). Older Ansible versions are known to miss IPv6 addresses
in interface facts. This may affect redeploys, reconfigures and
upgrades which run after VIP address is assigned.
See: https://github.com/ansible/ansible/issues/63227

Bifrost Train does not support IPv6 deployments.
See: https://storyboard.openstack.org/#!/story/2006689

Change-Id: Ia34e6916ea4f99e9522cd2ddde03a0a4776f7e2c
Implements: blueprint ipv6-control-plane
Signed-off-by: Radosław Piliszek <radoslaw.piliszek@gmail.com>
2019-10-16 10:24:35 +02:00

156 lines
5.6 KiB
YAML

---
- hosts: all
any_errors_fatal: true
vars:
logs_dir: "/tmp/logs"
roles:
- bindep
- multi-node-firewall
tasks:
# We have had cases where the nodepool private IP address is not assigned,
# which causes hard to diagnose errors later on. Catch it early.
- name: Assert that the nodepool private IPv4 address is assigned
assert:
that: nodepool.private_ipv4 in ansible_all_ipv4_addresses
fail_msg: >-
The nodepool private IP address {{ nodepool.private_ipv4 }} is not assigned
- name: Install dbus for debian system
apt: name=dbus
when:
- ansible_os_family == 'Debian'
become: true
- name: Ensure /tmp/logs/ dir
file:
path: "{{ logs_dir }}"
state: "directory"
- name: Ensure node directories
file:
path: "{{ logs_dir }}/{{ item }}"
state: "directory"
mode: 0777
with_items:
- "docker_logs"
- "kolla_configs"
- "system_logs"
- "kolla"
- "ansible"
- name: set new hostname based on ansible inventory file
hostname:
name: "{{ inventory_hostname }}"
become: true
# NOTE(yoctozepto): start VXLAN interface config
- name: Set VXLAN interface facts
set_fact:
api_interface_address: "{{ api_network_prefix }}{{ groups['all'].index(inventory_hostname) + 1 }}"
api_interface_tunnel_vni: 10001
tunnel_local_address: "{{ nodepool.private_ipv4 }}"
# NOTE(yoctozepto): CI VXLAN must use a different port than neutron-openvswitch-agent
# which defaults to 4789 (the default is used in CI)
# hence using port 4790
- name: Create VXLAN interface
become: true
command: ip link add {{ api_interface_name }} type vxlan id {{ api_interface_tunnel_vni }} local {{ tunnel_local_address }} dstport 4790
- name: Set VXLAN interface MTU
become: true
vars:
# Find the parent interface
parent_interface: >-
{{ ansible_interfaces |
map('extract', ansible_facts) |
selectattr('ipv4.address', 'defined') |
selectattr('ipv4.address', 'equalto', tunnel_local_address) |
first }}
# Allow 50 bytes overhead for VXLAN headers.
mtu: "{{ parent_interface.mtu | int - 50 }}"
command: ip link set {{ api_interface_name }} mtu {{ mtu }}
# emulate BUM by multiplicating traffic to unicast targets
- name: Add fdb entries for BUM traffic
become: true
vars:
dest_ip: "{{ hostvars[item].tunnel_local_address }}"
command: bridge fdb append 00:00:00:00:00:00 dev {{ api_interface_name }} dst {{ dest_ip }}
with_inventory_hostnames: all
when: item != inventory_hostname
- name: Add IPv4 address for VXLAN network
become: true
vars:
api_network_cidr: "{{ api_interface_address }}/{{ api_network_prefix_length }}"
# NOTE(yoctozepto): we have to compute and explicitly set the broadcast address,
# otherwise bifrost fails its pre-bootstrap sanity checks due to missing
# broadcast address as ansible picks up scope ('global') as the interface's
# broadcast address which fails checks logic
api_network_broadcast_address: "{{ api_network_cidr | ipaddr('broadcast') }}"
command: ip address add {{ api_network_cidr }} broadcast {{ api_network_broadcast_address }} dev {{ api_interface_name }}
when: address_family == 'ipv4'
# NOTE(yoctozepto): IPv6 has no broadcast address, let's not create confusion by setting it
- name: Add IPv6 address for VXLAN network
become: true
command: ip address add {{ api_interface_address }}/{{ api_network_prefix_length }} dev {{ api_interface_name }}
when: address_family == 'ipv6'
- name: Accept traffic on the VXLAN network (IN)
become: true
iptables:
state: present
action: insert
chain: INPUT
ip_version: "{{ address_family }}"
in_interface: "{{ api_interface_name }}"
jump: ACCEPT
# NOTE(yoctozepto): the default policy is ACCEPT but it is nicer to get statistics
- name: Accept traffic on the VXLAN network (OUT)
become: true
iptables:
state: present
action: insert
chain: OUTPUT
ip_version: "{{ address_family }}"
out_interface: "{{ api_interface_name }}"
jump: ACCEPT
- name: Bring VXLAN interface up
become: true
command: ip link set {{ api_interface_name }} up
# NOTE(yoctozepto): IPv6 DAD may delay proper address assignment
# this task will wait until DAD is done and addresses are no longer tentative
# we assign addresses uniquely so DAD can only move it to preferred
# hence we only check whether it's no longer tentative
- name: Ensure IPv6 addresses on VXLAN are no longer tentative
become: true
command: ip -o address show tentative dev {{ api_interface_name }}
register: tentative_addresses
until: tentative_addresses.stdout == ''
retries: 30
delay: 2
when:
- address_family == 'ipv6'
- name: Ping across VXLAN
vars:
ping_command: "{{ 'ping' if address_family == 'ipv4' else 'ping6' }}"
command: "{{ ping_command }} -c1 {{ hostvars[item].api_interface_address }}"
with_inventory_hostnames: all
# NOTE(yoctozepto): CentOS 7 image uses myhostname plugin for NSS
# which creates issues with IPv6-only deployment by providing
# an IPv4 address for the current hostname (affects rabbitmq)
- name: Disable myhostname NSS plugin
become: true
replace:
path: /etc/nsswitch.conf
regexp: ' myhostname'
replace: ''