Tungsten Fabric入门宝典系列文章，来自技术大牛倾囊相授的实践经验，由TF中文社区为您编译呈现，旨在帮助新手深入理解TF的运行、安装、集成、调试等全流程。如果您有相关经验或疑问，欢迎与我们互动，并与社区极客们进一步交流。更多TF技术文章，请点击公号底部按钮＞学习＞文章合集。
作者：Tatsuya Naganawa 译者：TF编译组
（编者按：在上一篇文章中，我们聊了关于TF组件安装中的HA行为、多NIC安装、集群规模等问题，以及如何安装具有Tungsten Fabric CNI的Kubernetes HA环境。）
OpenStack
Openstack HA的安装可以由anstack-deployer直接完成。
对于此示例的设置，我使用了5个EC2实例（AMI相同，ami-3185744e），每个实例具有2个vCPU、8 GB内存、20 GB磁盘空间。VPC的CIDR为172.31.0.0/16。
yum -y install epel-release
yum -y install git ansible-2.4.2.0
ssh-keygen
cd .ssh/
cat id_rsa.pub >> authorized_keys
cd
git clone http://github.com/Juniper/contrail-ansible-deployer
cd contrail-ansible-deployer
vi config/instances.yaml
(replace contents with this)
provider_config:
bms:
ssh_user: root
ssh_public_key: /root/.ssh/id_rsa.pub
ssh_private_key: /root/.ssh/id_rsa
domainsuffix: local
ntpserver: 0.centos.pool.ntp.org
instances:
bms1:
provider: bms
ip: 172.31.6.90 # controller1's ip
roles:
config_database:
config:
control:
analytics:
webui:
openstack:
bms2:
provider: bms
ip: 172.31.25.90 # controller2's ip
roles:
config_database:
config:
control:
analytics:
webui:
openstack:
bms3:
provider: bms
ip: 172.31.31.242 # controller3's ip
roles:
config_database:
config:
control:
analytics:
webui:
openstack:
bms11:
provider: bms
ip: 172.31.42.209 # compute1's ip
roles:
vrouter:
openstack_compute:
bms12:
provider: bms
ip: 172.31.15.199 # compute2's ip
roles:
vrouter:
openstack_compute:
contrail_configuration:
RABBITMQ_NODE_PORT: 5673
AUTH_MODE: keystone
KEYSTONE_AUTH_URL_VERSION: /v3
JVM_EXTRA_OPTS: "-Xms128m -Xmx1g"
kolla_config:
kolla_globals:
kolla_internal_vip_address: 172.31.0.11 ## kolla-ansible will deploy haproxy to serve HA vip
kolla_passwords:
keystone_admin_password: contrail123 # admin user's password
global_configuration:
## if previously described AMI is used, it uses cloud-init packages whose rpm dependency is not compatible with ansible-deployer in R5.1 and later. To workaroud this, I used these commands.
yum -y remove PyYAML python-requests
easy_install pip
pip install PyYAML requests
pip install ansible
ansible-playbook -e orchestrator=openstack -i inventory/ playbooks/configure_instances.yml
- it takes about 10 minutes
ansible-playbook -e orchestrator=openstack -i inventory/ playbooks/install_openstack.yml
- it takes about 40 minutes
ansible-playbook -e orchestrator=openstack -i inventory/ playbooks/install_contrail.yml
- it takes about 20 minutes
[root@ip-172-31-6-90 ~]# contrail-status
Pod Service Original Name State Id Status
redis contrail-external-redis running 23ef79b48ae8 Up 41 minutes
analytics api contrail-analytics-api running 3139f5fd9256 Up 36 minutes
analytics collector contrail-analytics-collector running 89c9e02fb551 Up 36 minutes
analytics nodemgr contrail-nodemgr running 5eecb461f95c Up 36 minutes
config api contrail-controller-config-api running fb0dc55f76c7 Up 39 minutes
config device-manager contrail-controller-config-devicemgr running 8dbff58776a2 Up 39 minutes
config nodemgr contrail-nodemgr running b64af838545d Up 39 minutes
config schema contrail-controller-config-schema running 83e0acf17e39 Up 39 minutes
config svc-monitor contrail-controller-config-svcmonitor running 623e17e8e74e Up 39 minutes
config-database cassandra contrail-external-cassandra running db30d874dce3 Up 40 minutes
config-database nodemgr contrail-nodemgr running 590463f627f6 Up 38 minutes
config-database rabbitmq contrail-external-rabbitmq running 712ee26dda64 Up 40 minutes
config-database zookeeper contrail-external-zookeeper running 46dbdec00e46 Up 40 minutes
control control contrail-controller-control-control running 3e0e653d1588 Up 37 minutes
control dns contrail-controller-control-dns running 2cebc37c18cf Up 37 minutes
control named contrail-controller-control-named running 112bd2d8ed5f Up 37 minutes
control nodemgr contrail-nodemgr running f2e0fdc4bfb2 Up 37 minutes
device-manager dnsmasq contrail-external-dnsmasq running f84b45234d70 Up 39 minutes
webui job contrail-controller-webui-job running 3dece86513a1 Up 38 minutes
webui web contrail-controller-webui-web running 408c772b1628 Up 38 minutes
== Contrail control ==
control: active
nodemgr: active
named: active
dns: active
== Contrail config-database ==
nodemgr: initializing (Disk for DB is too low. )
zookeeper: active
rabbitmq: active
cassandra: active
== Contrail analytics ==
nodemgr: active
api: active
collector: active
== Contrail webui ==
web: active
job: active
== Contrail device-manager ==
== Contrail config ==
svc-monitor: backup
nodemgr: active
device-manager: backup
api: active
schema: backup
[root@ip-172-31-25-90 ~]# contrail-status
Pod Service Original Name State Id Status
redis contrail-external-redis running 1ed7e967085e Up 41 minutes
analytics api contrail-analytics-api running 7392ea345e83 Up 36 minutes
analytics collector contrail-analytics-collector running 82332a53a566 Up 36 minutes
analytics nodemgr contrail-nodemgr running 89141bb180cd Up 36 minutes
config api contrail-controller-config-api running b2af8bc8a6d7 Up 38 minutes
config device-manager contrail-controller-config-devicemgr running d8ed77431dfa Up 39 minutes
config nodemgr contrail-nodemgr running 8c7f3d5f05e4 Up 39 minutes
config schema contrail-controller-config-schema running 4a6099aaea2a Up 39 minutes
config svc-monitor contrail-controller-config-svcmonitor running 3a3e6d37b30e Up 39 minutes
config-database cassandra contrail-external-cassandra running 0b05e121c017 Up 40 minutes
config-database nodemgr contrail-nodemgr running fb4857fe16c1 Up 39 minutes
config-database rabbitmq contrail-external-rabbitmq running a8137277a40f Up 40 minutes
config-database zookeeper contrail-external-zookeeper running 9571f4d9fde2 Up 40 minutes
control control contrail-controller-control-control running 5460dc02cc03 Up 37 minutes
control dns contrail-controller-control-dns running 17b27877ef6e Up 37 minutes
control named contrail-controller-control-named running cdbe1bae4c40 Up 37 minutes
control nodemgr contrail-nodemgr running cb36c2b4625a Up 37 minutes
device-manager dnsmasq contrail-external-dnsmasq running dd9002e6f58d Up 39 minutes
webui job contrail-controller-webui-job running 60dc895d439e Up 38 minutes
webui web contrail-controller-webui-web running 3ddfb5e2e851 Up 38 minutes
== Contrail control ==
control: active
nodemgr: active
named: active
dns: active
== Contrail config-database ==
nodemgr: initializing (Disk for DB is too low. )
zookeeper: active
rabbitmq: active
cassandra: active
== Contrail analytics ==
nodemgr: active
api: active
collector: active
== Contrail webui ==
web: active
job: active
== Contrail device-manager ==
== Contrail config ==
svc-monitor: backup
nodemgr: active
device-manager: active
api: active
schema: backup
[root@ip-172-31-31-242 ~]# contrail-status
Pod Service Original Name State Id Status
redis contrail-external-redis running 172e35daca5a Up 42 minutes
analytics api contrail-analytics-api running 2edf90837a43 Up 36 minutes
analytics collector contrail-analytics-collector running 812d4c190841 Up 36 minutes
analytics nodemgr contrail-nodemgr running d0eafce0d49d Up 36 minutes
config api contrail-controller-config-api running 7819c7792960 Up 39 minutes
config device-manager contrail-controller-config-devicemgr running c22addf8f1f1 Up 38 minutes
config nodemgr contrail-nodemgr running bd742928f26e Up 39 minutes
config schema contrail-controller-config-schema running 8ad72d0a2c12 Up 39 minutes
config svc-monitor contrail-controller-config-svcmonitor running 86283bfc21dc Up 39 minutes
config-database cassandra contrail-external-cassandra running 315d17494665 Up 41 minutes
config-database nodemgr contrail-nodemgr running a78521b2b940 Up 39 minutes
config-database rabbitmq contrail-external-rabbitmq running dfefb054808b Up 41 minutes
config-database zookeeper contrail-external-zookeeper running a16d1a2d259b Up 41 minutes
control control contrail-controller-control-control running bc9ecb41131c Up 37 minutes
control dns contrail-controller-control-dns running beff8cf11fdd Up 37 minutes
control named contrail-controller-control-named running 2322d5598a24 Up 37 minutes
control nodemgr contrail-nodemgr running 32b611d85d19 Up 37 minutes
device-manager dnsmasq contrail-external-dnsmasq running a0b3dd0ad254 Up 39 minutes
webui job contrail-controller-webui-job running 257721b46207 Up 38 minutes
webui web contrail-controller-webui-web running c2e7b95e7321 Up 38 minutes
== Contrail control ==
control: active
nodemgr: active
named: active
dns: active
== Contrail config-database ==
nodemgr: initializing (Disk for DB is too low. )
zookeeper: active
rabbitmq: active
cassandra: active
== Contrail analytics ==
nodemgr: active
api: active
collector: active
== Contrail webui ==
web: active
job: active
== Contrail device-manager ==
== Contrail config ==
svc-monitor: active
nodemgr: active
device-manager: backup
api: active
schema: active
[root@ip-172-31-42-209 ~]# contrail-status
Pod Service Original Name State Id Status
vrouter agent contrail-vrouter-agent running a17883037f12 Up 36 minutes
vrouter nodemgr contrail-nodemgr running 6dc2258ac4f6 Up 36 minutes
vrouter kernel module is PRESENT
== Contrail vrouter ==
nodemgr: active
agent: active
[root@ip-172-31-15-199 ~]# contrail-status
Pod Service Original Name State Id Status
vrouter agent contrail-vrouter-agent running a1e7767b3302 Up 36 minutes
vrouter nodemgr contrail-nodemgr running 40d5613fec21 Up 36 minutes
vrouter kernel module is PRESENT
== Contrail vrouter ==
nodemgr: active
agent: active然后，可以使用openstack命令创建实例。
docker cp /etc/kolla/kolla-toolbox/admin-openrc.sh kolla_toolbox:/var/tmp
docker exec -it kolla_toolbox bash
source /var/tmp/admin-openrc.sh
cd /var/tmp
curl -O -L http://download.cirros-cloud.net/0.4.0/cirros-0.4.0-x86_64-disk.img
openstack image create cirros --disk-format qcow2 --public --container-format bare --file cirros-0.4.0-x86_64-disk.img
openstack flavor create --ram 512 --disk 1 --vcpus 1 m1.tiny
openstack network create testvn
openstack subnet create --subnet-range 192.168.100.0/24 --network testvn subnet1
NET_ID=`openstack network list | grep testvn | awk -F '|' '{print $2}' | tr -d ' '`
openstack server create --flavor m1.tiny --image cirros --nic net-id=${NET_ID} vm1
openstack server create --flavor m1.tiny --image cirros --nic net-id=${NET_ID} vm2
exit
(on compute nodes)
ip route ## check metadata ip of two instances
ssh [email protected]
ping 192.168.100.4
(kolla-toolbox)[ansible@ip-172-31-6-90 /]$ openstack server list
+--------------------------------------+------+--------+----------------------+--------+---------+
| ID | Name | Status | Networks | Image | Flavor |
+--------------------------------------+------+--------+----------------------+--------+---------+
| 9d66f0ed-d7d5-4a53-983d-dfba0385bd22 | vm2 | ACTIVE | testvn=192.168.100.4 | cirros | m1.tiny |
| 6595b4c1-1e6f-4f02-8f66-83b6355065b2 | vm1 | ACTIVE | testvn=192.168.100.3 | cirros | m1.tiny |
+--------------------------------------+------+--------+----------------------+--------+---------+
(kolla-toolbox)[ansible@ip-172-31-6-90 /]$
[root@ip-172-31-42-209 ~]# ip route
default via 172.31.32.1 dev vhost0
169.254.0.1 dev vhost0 proto 109 scope link
169.254.0.3 dev vhost0 proto 109 scope link
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1
172.31.32.0/20 dev vhost0 proto kernel scope link src 172.31.42.209
[root@ip-172-31-42-209 ~]# ssh [email protected]
[email protected]'s password:
$ ip -o a
1: lo: mtu 65536 qdisc noqueue qlen 1\ link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
1: lo inet 127.0.0.1/8 scope host lo\ valid_lft forever preferred_lft forever
1: lo inet6 ::1/128 scope host \ valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast qlen 1000\ link/ether 02:79:59:ea:d4:17 brd ff:ff:ff:ff:ff:ff
2: eth0 inet 192.168.100.3/24 brd 192.168.100.255 scope global eth0\ valid_lft forever preferred_lft forever
2: eth0 inet6 fe80::79:59ff:feea:d417/64 scope link \ valid_lft forever preferred_lft forever
$
$ ping 192.168.100.4
PING 192.168.100.4 (192.168.100.4): 56 data bytes
64 bytes from 192.168.100.4: seq=0 ttl=64 time=13.876 ms
64 bytes from 192.168.100.4: seq=1 ttl=64 time=2.417 ms
64 bytes from 192.168.100.4: seq=2 ttl=64 time=2.375 ms
^C
--- 192.168.100.4 ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
round-trip min/avg/max = 2.375/6.222/13.876 ms
$
$
[root@ip-172-31-15-199 ~]# ip route
default via 172.31.0.1 dev vhost0
169.254.0.1 dev vhost0 proto 109 scope link
169.254.0.3 dev vhost0 proto 109 scope link
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1
172.31.0.0/20 dev vhost0 proto kernel scope link src 172.31.15.199
[root@ip-172-31-15-199 ~]# ssh [email protected]
[email protected]'s password:
$ ip -o a
1: lo: mtu 65536 qdisc noqueue qlen 1\ link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
1: lo inet 127.0.0.1/8 scope host lo\ valid_lft forever preferred_lft forever
1: lo inet6 ::1/128 scope host \ valid_lft forever preferred_lft forever
2: eth0: mtu 1500 qdisc pfifo_fast qlen 1000\ link/ether 02:08:e6:0d:1e:3b brd ff:ff:ff:ff:ff:ff
2: eth0 inet 192.168.100.4/24 brd 192.168.100.255 scope global eth0\ valid_lft forever preferred_lft forever
2: eth0 inet6 fe80::8:e6ff:fe0d:1e3b/64 scope link \ valid_lft forever preferred_lft forever
$ 注意：如果计算节点不支持kvm，则可能需要添加此设置。
vi /etc/kolla/nova-compute/nova.conf
(add them in [libvirt] section)
virt_type=qemu
cpu_mode=none
docker restart nova_compute注意：如果是使用AWS，则还需要从EC2实例的右键菜单中设置网络>管理IP地址，以允许从其它节点访问haproxy VIP。
这样，控制器之间的HA和2个计算节点之间的overlay都完全配置好了！
有些要点本文档没有覆盖，例如某些控制节点宕机时的行为，或在计算节点之间执行实时迁移。上次我尝试实时迁移时，发现有大约1秒钟的报文丢包，但是请注意自己的设置，因为有很多要注意的地方（例如前缀将在实时迁移完成后更新）。
查看每个控制器的邻居状态和路由表，您会发现它们之间的有趣之处。
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py ctr nei
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| peer | peer_address | peer_asn | encoding | peer_type | state | send_state | flap_count | flap_time |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| ip-172-31-25-90.local | 172.31.25.90 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-31-242.local | 172.31.31.242 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-42-209.local | 172.31.42.209 | 0 | XMPP | internal | Established | in sync | 0 | n/a |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.25.90 ctr nei
Introspect Host: 172.31.25.90
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| peer | peer_address | peer_asn | encoding | peer_type | state | send_state | flap_count | flap_time |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| ip-172-31-31-242.local | 172.31.31.242 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-6-90.local | 172.31.6.90 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-15-199.local | 172.31.15.199 | 0 | XMPP | internal | Established | in sync | 0 | n/a |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
[root@ip-172-31-6-90 ~]#
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.31.242 ctr nei
Introspect Host: 172.31.31.242
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| peer | peer_address | peer_asn | encoding | peer_type | state | send_state | flap_count | flap_time |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
| ip-172-31-25-90.local | 172.31.25.90 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-6-90.local | 172.31.6.90 | 64512 | BGP | internal | Established | in sync | 0 | n/a |
| ip-172-31-15-199.local | 172.31.15.199 | 0 | XMPP | internal | Established | in sync | 0 | n/a |
| ip-172-31-42-209.local | 172.31.42.209 | 0 | XMPP | internal | Established | in sync | 0 | n/a |
+------------------------+---------------+----------+----------+-----------+-------------+------------+------------+-----------+
[root@ip-172-31-6-90 ~]#
[root@ip-172-31-6-90 ~]#
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py ctr route summary
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| name | prefixes | paths | primary_paths | secondary_paths | infeasible_paths |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| default-domain:admin:testvn:testvn.inet.0 | 2 | 4 | 1 | 3 | 0 |
| default-domain:default- | 0 | 0 | 0 | 0 | 0 |
| project:__link_local__:__link_local__.inet.0 | | | | | |
| default-domain:default-project:dci- | 0 | 0 | 0 | 0 | 0 |
| network:__default__.inet.0 | | | | | |
| default-domain:default-project:dci-network:dci- | 0 | 0 | 0 | 0 | 0 |
| network.inet.0 | | | | | |
| default-domain:default-project:default-virtual- | 0 | 0 | 0 | 0 | 0 |
| network:default-virtual-network.inet.0 | | | | | |
| inet.0 | 0 | 0 | 0 | 0 | 0 |
| default-domain:default-project:ip-fabric:ip- | 1 | 1 | 1 | 0 | 0 |
| fabric.inet.0 | | | | | |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.25.90 ctr route summary
Introspect Host: 172.31.25.90
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| name | prefixes | paths | primary_paths | secondary_paths | infeasible_paths |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| default-domain:admin:testvn:testvn.inet.0 | 2 | 4 | 1 | 3 | 0 |
| default-domain:default- | 0 | 0 | 0 | 0 | 0 |
| project:__link_local__:__link_local__.inet.0 | | | | | |
| default-domain:default-project:dci- | 0 | 0 | 0 | 0 | 0 |
| network:__default__.inet.0 | | | | | |
| default-domain:default-project:dci-network:dci- | 0 | 0 | 0 | 0 | 0 |
| network.inet.0 | | | | | |
| default-domain:default-project:default-virtual- | 0 | 0 | 0 | 0 | 0 |
| network:default-virtual-network.inet.0 | | | | | |
| inet.0 | 0 | 0 | 0 | 0 | 0 |
| default-domain:default-project:ip-fabric:ip- | 1 | 1 | 1 | 0 | 0 |
| fabric.inet.0 | | | | | |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.31.242 ctr route summary
Introspect Host: 172.31.31.242
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| name | prefixes | paths | primary_paths | secondary_paths | infeasible_paths |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
| default-domain:admin:testvn:testvn.inet.0 | 2 | 4 | 2 | 2 | 0 |
| default-domain:default- | 0 | 0 | 0 | 0 | 0 |
| project:__link_local__:__link_local__.inet.0 | | | | | |
| default-domain:default-project:dci- | 0 | 0 | 0 | 0 | 0 |
| network:__default__.inet.0 | | | | | |
| default-domain:default-project:dci-network:dci- | 0 | 0 | 0 | 0 | 0 |
| network.inet.0 | | | | | |
| default-domain:default-project:default-virtual- | 0 | 0 | 0 | 0 | 0 |
| network:default-virtual-network.inet.0 | | | | | |
| inet.0 | 0 | 0 | 0 | 0 | 0 |
| default-domain:default-project:ip-fabric:ip- | 2 | 2 | 2 | 0 | 0 |
| fabric.inet.0 | | | | | |
+----------------------------------------------------+----------+-------+---------------+-----------------+------------------+
[root@ip-172-31-6-90 ~]#
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.31.242 ctr route show 192.168.100.3
Introspect Host: 172.31.31.242
default-domain:admin:testvn:testvn.inet.0: 2 destinations, 4 routes (2 primary, 2 secondary, 0 infeasible)
192.168.100.3/32, age: 0:01:18.234010, last_modified: 2019-Apr-27 14:03:19.075046
[XMPP (interface)|ip-172-31-42-209.local] age: 0:01:18.239011, localpref: 200, nh: 172.31.42.209, encap: ['gre', 'udp'], label: 25, AS path: None
[BGP|172.31.6.90] age: 0:01:18.230559, localpref: 200, nh: 172.31.42.209, encap: ['gre', 'udp'], label: 25, AS path: None
[root@ip-172-31-6-90 ~]# ./contrail-introspect-cli/ist.py --host 172.31.31.242 ctr route show 192.168.100.4
Introspect Host: 172.31.31.242
default-domain:admin:testvn:testvn.inet.0: 2 destinations, 4 routes (2 primary, 2 secondary, 0 infeasible)
192.168.100.4/32, age: 0:00:52.035230, last_modified: 2019-Apr-27 14:03:47.460835
[XMPP (interface)|ip-172-31-15-199.local] age: 0:00:52.039485, localpref: 200, nh: 172.31.15.199, encap: ['gre', 'udp'], label: 25, AS path: None
[BGP|172.31.25.90] age: 0:00:51.996464, localpref: 200, nh: 172.31.15.199, encap: ['gre', 'udp'], label: 25, AS path: None
[root@ip-172-31-6-90 ~]# 由于vRouter始终具有2个XMPP连接，所以当存在3个控制器时，XMPP的连接状态在控制器之间是不同的，并且它们之间的路由表也可能有所不同。考虑到路由目标过滤，如果某些控制器未从XMPP接收到某些特定的路由目标，则它们甚至可能具有完全不同的路由表。
这就是由于Tungsten Fabric的横向扩展行为带来的结果。
有关ansible-ployer的更详细的配置（包括多NIC示例），您可以看看这些文档。
-
https://github.com/Juniper/contrail-ansible-deployer/wiki/Contrail-with-Openstack-Kolla
-
https://github.com/Juniper/contrail-ansible-deployer/wiki/Configuration-Sample-for-Multi-Node-Openstack-HA-and-Contrail-(single-interface)
- https://github.com/Juniper/contrail-ansible-deployer/wiki/Configuration-Sample-for-Multi-Node-Openstack-HA-and-Contrail-(multi-interface)
vCenter
Tungsten Fabric可以很好地与vCenter集成在一起，我们在“编排器集成”的部分有所描述。
要尝试此功能，可以按照以下说明进行操作。
- https://github.com/Juniper/contrail-ansible-deployer/blob/master/README_vcenter.m
- 请注意vRouterVM的OVF文件目前还未公开提供
由于Tungsten Fabric的HA行为与Kubernetes和OpenStack安装中的行为完全相同，因此我将不对其进行详细描述。
对于编排器端的HA，vCenter HA是完全可用的。
- 我尚未尝试过这种组合，但是由于vCenter HA将使用与vCenter服务相同的IP，因此我认为vcenter-plugin与vCenter HA一起使用的可能性很高。
多vCenter或跨vCenter（使用链接模式时）将是一个很有趣的主题。我将在附录部分进一步讨论它们。
容器标签的使用
容器注册表docker.io/opencontrailnight具有各种各样的标签。
- https://hub.docker.com/r/opencontrailnightly/contrail-nodemgr/tags/
关于在新的安装中选择什么样的标签，我有一些考虑。
我经常使用的是三个标签：latest，5.1.0-latest，5.0-latest。它们位于每个Tungsten Fabric分支的头部（master/R5.1/R5.0），并且每个分支中都包含各种错误修复。因此，您可以为用例选择其中的标签。如果需要R5.1中的新功能（例如可选的analytics组件），那么你可以选择5.1.0-latest标签。
由于latest是真正的开发分支，因此我不建议您将它们作为常规使用的标签，因为在某些情况下，此版本已被“破坏”以添加新的功能。
其它发行版本的分支更加稳定，因为在大多数情况下，它们只是进行了错误修复，尽管在创建新的分支后的某些特定时期内，发行版本的分支似乎也具有新的功能。
要指定标签，您可以使用这些参数，并且当针对ansible-deployer和conutil-container-builder输入git clone时，还需要指定同一分支。
(ansible-deployer)
git clone -b R5.1 http://github.com/Juniper/contrail-ansible-deployer
contrail_configuration:
CONTRAIL_CONTAINER_TAG: 5.1.0-latest
(kubeadm)
git clone -b R5.1 https://github.com/Juniper/contrail-container-builder.git
common.env in contrail-container-builder repo
CONTRAIL_CONTAINER_TAG: 5.1.0-latest需要注意的一点是，由于与OpenStack一起使用的容器（例如nova-init、neutron-init、heat-init等）具有OpenStack发行版的版本依赖性，因此标记可能需要更改为5.1.0-latest-queens、5.1.0-latest-rocky等。
由于安装了一些带有特定版本的OpenStack模块，因此，如果标签不同，则OpenStack容器将无法正常工作。
Tungsten Fabric入门宝典系列文章——
1.首次启动和运行指南
- TF组件的七种“武器”
- 编排器集成
-
关于安装的那些事（上）
Tungsten Fabric 架构解析系列文章——
第一篇：TF主要特点和用例
第二篇：TF怎么运作
第三篇：详解vRouter体系结构
第四篇：TF的服务链
第五篇：vRouter的部署选项
第六篇：TF如何收集、分析、部署？
第七篇：TF如何编排
第八篇：TF支持API一览
第九篇：TF如何连接到物理网络
第十篇：TF基于应用程序的安全策略
