openstack octavia 简介以及手工安装过程

openstack octavia 是 openstack lbaas的支持的一种后台程序，提供为虚拟机流量的负载均衡。实质是类似于trove，调用 nove 以及neutron的api生成一台安装好haproxy和keepalived软件的虚拟机，并连接到目标网路。octavia共有4个组件 housekeeping,worker,api,health-manager，octavia agent。api作用就不详细说了。worker：主要作用是和nova，neutron等组件通信，用于虚拟机调度以及把对于虚拟机操作的指令下发给octavia agent。housekeeping：查看octavia/controller/housekeeping/house_keeping.py得知其三个功能点：SpareAmphora,DatabaseCleanup,CertRotation。依次是清理虚拟机的池子，清理过期数据库，更新证书。health-manager：检查虚拟机状态，和虚拟机中的octavia agent通信，来更新各个组件的状态。octavia agent 位于虚拟机内部：对下是接受指令操作下层的haproxy软件，对上是和health-manager通信汇报各种情况。可以参考博文http://lingxiankong.github.io/blog/2016/03/30/octavia/?utm_source=tuicool&utm_medium=referral

写的比我更详细一点

目前官方不提供安装文档。谷歌了下似乎也没人写过具体的安装步骤，只推荐用devstack来进行安装。本人尝试根据devstack的安装脚本总结了下安装octavia的步骤，验证是成功的，不当之处请各位指正。

一 安装

1、创建数据库

1 2	mysql> CREATE DATABASE octavia; mysql> GRANT ALL PRIVILEGES ON octavia.* TO  'octavia' @ 'localhost'   IDENTIFIED BY  'OCTAVIA_DBPASS' ;mysql> GRANT ALL PRIVILEGES ON octavia.* TO  'octavia' @ '%'  \  IDENTIFIED BY  'OCTAVIA_DBPASS' ;

2 创建用户 角色 endpoint

1 2 3 4 5

openstack user create --domain default --password-prompt octavia openstack role add --project service --user cinder admin openstack endpoint create octavia public http: //10 .1.65.58:9876/ --region RegionOne  openstack endpoint create octavia admin http: //10 .1.65.58:9876/ --region RegionOne  openstack endpoint create octavia internal http: //10 .1.65.58:9876/ --region RegionOne

3 安装软件包

1	yum  install  openstack-octavia-worker openstack-octavia-api python-octavia openstack-octavia openstack-octavia openstack-octavia

4 导入镜像 镜像是从devstack 生成的系统中导出来的

1	openstack  image create amphora-x64-haproxy --public --container- format =bare --disk- format  qcow2

5 创建管理网络，并在主机创建ovs端口，使octavia-worker，octavia-housekeeping，octavia-health-manager能和生成的虚拟机实例通讯

 5.1 生成管理网络，网段

1 2	openstack network create lb-mgmt-net openstack subnet create --subnet-range 192.168.0.0 /24  --allocation-pool       start=192.168.0.2,end=192.168.0.200 --network lb-mgmt-net lb-mgmt-subnet

 5.2 生成管理端口防火墙规则 

5555端口是管理网络，考虑到octavia组件尚不成熟，开启了22端口，镜像本身也是开启了22端口，这点吐槽下trove，同样是不成熟的模块，默认不开启22端口，还得去改源码。

1 2 3

openstack security group create lb-mgmt-sec-grp openstack security group rule create --protocol udp --dst-port 5555 lb-health-mgr-sec-grp openstack security group rule create --protocol tcp --dst-port 22 lb-mgmt-sec-grp

 5.3 在管理网络创建一个端口用于连接宿主机中的octavia health_manager

1	neutron port-create --name octavia-health-manager-standalone-listen-port --security-group lb-health-mgr-sec-grp --device-owner Octavia:health-mgr --binding:host_id=controller lb-mgmt-net

 5.4 创建宿主机的ovs端口 并连接至5.1生成的网络

1	ovs-vsctl  --may-exist add-port br-int o-hm0 --  set  Interface o-hm0  type =internal --  set  Interface o-hm0 external-ids:iface-status=active --  set  Interface o-hm0 external-ids:attached-mac=fa:16:3e:6f:9f:9a --  set  Interface o-hm0 external-ids:iface- id =457e4953-b2d6-49ee-908b-2991506602b2

其中iface-id 和attached-mac 为 5.3生成的port的 属性

1	ip link  set  dev o-hm0 address fa:16:3e:6f:9f:9a

 5.5 在宿主机上创建dhcp （为啥不用传统的dnsmasq呢？）

1	dhclient - v  o-hm0 -cf  /etc/octavia/dhcp/dhclient .conf

6 配置修改，和其他openstack组件设置差不多

  6.1 设置数据库

1 2	[database] connection = mysql+pymysql: //octavia :OCTAVIA_DBPASS@controller /octavia

  6.2 设置消息队列

1 2 3 4

[oslo_messaging_rabbit] rabbit_host = controller rabbit_userid = openstack rabbit_password = RABBIT_PASS

6.3 设置 keystone的认证信息

1 2 3 4 5 6

[keystone_authtoken] auth_version = 2 admin_password = OCTAVIA_PASS admin_tenant_name = octavia admin_user = octavia auth_uri = http: //controller :5000 /v2 .0

6.4 设置health_manager组件监听地址，此ip地址等于5.3中创建的io地址

1 2 3 4

[health_manager] bind_port = 5555 bind_ip = 192.168.0.7 controller_ip_port_list = 192.168.0.7:5555

6.5 设置和虚拟机通信的 公钥私钥

1 2 3 4 5 6 7 8

[haproxy_amphora] server_ca =  /etc/octavia/certs/ca_01 .pem client_cert =  /etc/octavia/certs/client .pem key_path =  /etc/octavia/ . ssh /octavia_ssh_key base_path =  /var/lib/octavia base_cert_dir =  /var/lib/octavia/certs connection_max_retries = 1500 connection_retry_interval = 1

6.6 设置 用于开启虚拟机实例的信息

1 2 3 4 5 6 7 8 9 10 11 12

[controller_worker] amp_boot_network_list = 826be4f4-a23d-4c5c-bff5-7739936fac76  # 步骤5.1中生成的id amp_image_tag = amphora  # 步骤4 中已经定义这个metadata amp_secgroup_list = d949202b-ba09-4003-962f-746ae75809f7  # 步骤5.2 生成的安全组id amp_flavor_id = dd49b3d5-4693-4407-a76e-2ca95e00a9ec amp_image_id = b23dda5f-210f-40e6-9c2c-c40e9daa661a  #步骤4中生成的镜像id amp_ssh_key_name = 155  # amp_active_wait_sec = 1 amp_active_retries = 100 network_driver = allowed_address_pairs_driver compute_driver = compute_nova_driver amphora_driver = amphora_haproxy_rest_driver

7 修改neutron配置

  7.1 修改 /etc/neutron/neutron.conf 增加lbaas服务

 

1	service_plugins = [existing service plugins],neutron_lbaas.services.loadbalancer.plugin.LoadBalancerPluginv2

7.2 在[service_providers] 章节 设置lbaas 的服务提供者为octavia

 

1	service_provider = LOADBALANCERV2:Octavia:neutron_lbaas.drivers.octavia.driver.OctaviaDriver:default

8 启动服务

如果之前 开启了 LBaaS v2 with an agent 服务 请关闭，并清理下neutron数据库下 lbaas_loadbalancers lbaas_loadbalancer_statistics 不然会报错

同步数据库

 

1	octavia-db-manage   upgrade  head

重启neutron 

1	systemctl restart neutron-server

启动octavia

1	systemctl restart  octavia-housekeeping  octavia-worker octavia-api octavia-health-manager

二 验证

9.1 创建loadbalancer

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

[root@controller ~] # neutron lbaas-loadbalancer-create --name test-lb-1 lbtest Created a new loadbalancer: +---------------------+--------------------------------------+ | Field               | Value                                | +---------------------+--------------------------------------+ | admin_state_up      | True                                 | | description         |                                      | |  id                   | 5af472bb-2068-4b96-bcb3-bef7ff7abc56 | | listeners           |                                      | | name                |  test -lb-1                            | | operating_status    | OFFLINE                              | | pools               |                                      | | provider            | octavia                              | | provisioning_status | PENDING_CREATE                       | | tenant_id           | 9a4b2de78c2d45cfbf6880dd34877f7b     | | vip_address         | 192.168.123.10                       | | vip_port_id         | d163b73c-258a-4e03-90ad-5db31cfe23ac | | vip_subnet_id       | 74aea53a-014a-4f9c-86f9-805a2a772a27 | +---------------------+--------------------------------------+

 9.2 查看虚拟机，值得注意的地方，loadbalancer的地址是vip，和虚拟机的地址是不相同的

1 2	[root@controller ~] # openstack server list |grep 82b59e85-29f2-46ce-ae0b-045b7fceb5ca | 82b59e85-29f2-46ce-ae0b-045b7fceb5ca | amphora-734da57c-e444-4b8e-a706-455230ae0803 | ACTIVE  | lbtest=192.168.123.9; lb-mgmt-net=192.168.0.6        | amphora-x64-haproxy 201610131607    |

 9.3 创建linstener

1	neutron lbaas-listener-create --name  test -lb-tcp --loadbalancer  test -lb-1 --protocol TCP  --protocol-port 22

 9.4 设置安全组

1	neutron port-update  --security-group default d163b73c-258a-4e03-90ad-5db31cfe23ac

 9.5 创建pool,新建三台虚拟机，并加入pool

1 2 3 4 5 6 7 8 9 10 11 12 13

openstack server create  --flavor m1.small --nic net- id =22525640-297e-40eb-bd77-0a9afd861f8c --image  "cirros for kvm raw"   --min 3 --max 3  test   [root@controller ~] # openstack server list |grep test- | d8dc22d4-e657-4c54-96f9-3a53ca67533d |  test -3                                       | ACTIVE  | lbtest=192.168.123.8                                 | cirros  for  kvm raw                  | | c7926665-84c5-48a5-9de5-5e15e71baa5d |  test -2                                       | ACTIVE  | lbtest=192.168.123.13                                | cirros  for  kvm raw                  | | fcf60c23-b799-4d08-a5a7-2b0fc9f1905e |  test -1                                       | ACTIVE  | lbtest=192.168.123.11                                | cirros  for  kvm raw                  |   neutron lbaas-pool-create   --name  test -lb-pool-tcp  --lb-algorithm ROUND_ROBIN --listener  test -lb-tcp --protocol TCP    for  i  in  {8,13,11} do neutron lbaas-member-create --subnet lbtest  --address 192.168.123.${i}  --protocol-port 22   test -lb-pool-tcp done

 9.6 验证

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

[root@controller ~] # >/root/.ssh/known_hosts;ip netns exec qrouter-4718cc34-68cc-47a7-9201-405d1fc09213 ssh cirros@192.168.123.10 "hostname" The authenticity of host  '192.168.123.10 (192.168.123.10)'  can't be established. RSA key fingerprint is 72:c4:11:41:53:51:f2:1b:b5:e6:1b:69:a8:c2:5b:d4. Are you sure you want to  continue  connecting ( yes /no )?  yes Warning: Permanently added  '192.168.123.10'  (RSA) to the list of known hosts. cirros@192.168.123.10's password:  test -3 [root@controller ~] # >/root/.ssh/known_hosts;ip netns exec qrouter-4718cc34-68cc-47a7-9201-405d1fc09213 ssh cirros@192.168.123.10 "hostname" The authenticity of host  '192.168.123.10 (192.168.123.10)'  can't be established. RSA key fingerprint is 3d:88:0f:4a:b1:77:c9:6a:fd:82:4d:31:0c:ca:82:d8. Are you sure you want to  continue  connecting ( yes /no )?  yes Warning: Permanently added  '192.168.123.10'  (RSA) to the list of known hosts. cirros@192.168.123.10's password:  test -1 [root@controller ~] # >/root/.ssh/known_hosts;ip netns exec qrouter-4718cc34-68cc-47a7-9201-405d1fc09213 ssh cirros@192.168.123.10 "hostname" The authenticity of host  '192.168.123.10 (192.168.123.10)'  can't be established. RSA key fingerprint is 1c:03:f0:f9:92:a7:0f:5d:9d:09:22:14:94:62:e4:c4. Are you sure you want to  continue  connecting ( yes /no )?  yes Warning: Permanently added  '192.168.123.10'  (RSA) to the list of known hosts. cirros@192.168.123.10's password:  test -2

三 过程分析

10.1 worker的相关操作

 创建 云主机实例，关联到管理网络：

1

REQ: curl -g -i -X POST http: //controller :8774 /v2 .1 /9a4b2de78c2d45cfbf6880dd34877f7b/servers  -H  "User-Agent: python-novaclient"  -H  "Content-Type: application/json"  -H  "Accept: application/json"  -H  "X-OpenStack-Nova-API-Version: 2.1"  -H  "X-Auth-Token: {SHA1}0f810ab0fdd5b92489f73a7f0988adfc9da4e517"  -d  '{"server": {"name": "amphora-4f22d55b-0680-4111-aef6-da98c9ccd1d4", "imageRef": "b23dda5f-210f-40e6-9c2c-c40e9daa661a", "key_name": "155", "flavorRef": "dd49b3d5-4693-4407-a76e-2ca95e00a9ec", "max_count": 1, "min_count": 1, "personality": [{"path": "/etc/octavia/amphora-agent.conf", "contents": ""}, {"path": "/etc/octavia/certs/client_ca.pem", "contents": "="}, {"path": "/etc/octavia/certs/server.pem", "contents": ""}], "networks": [{"uuid": "826be4f4-a23d-4c5c-bff5-7739936fac76"}], "security_groups": [{"name": "d949202b-ba09-4003-962f-746ae75809f7"}], "config_drive": true}}'  _http_log_request  /usr/lib/python2 .7 /site-packages/keystoneauth1/session .py:337

当检测到目标云主机的管理网络状态变为active后进行下一步

1 2 3

REQ: curl -g -i -X GET http: //controller :8774 /v2 .1 /9a4b2de78c2d45cfbf6880dd34877f7b/servers/d3c97360-56b2-4f75-b905-2ef83870a342/os-interface  -H  "User-Agent: python-novaclient"  -H  "Accept: application/json"  -H  "X-OpenStack-Nova-API-Version: 2.1"  -H  "X-Auth-Token: {SHA1}3f6ccac4cb8b70b06fb5e62b9db2272702d8ec67"  _http_log_request  /usr/lib/python2 .7 /site-packages/keystoneauth1/session .py:337 2016-10-17 12:06:30.041 29993 DEBUG novaclient.v2.client [-] RESP: [200] Content-Length: 286 Content-Type: application /json  Openstack-Api-Version: compute 2.1 X-Openstack-Nova-Api-Version: 2.1 Vary: OpenStack-API-Version, X-OpenStack-Nova-API-Version X-Compute-Request-Id: req-ccc07b37-e942-4a5b-a87a-b0e8d3887ba3 Date: Mon, 17 Oct 2016 04:06:30 GMT Connection: keep-alive RESP BODY: { "interfaceAttachments" : [{ "port_state" :  "ACTIVE" ,  "fixed_ips" : [{ "subnet_id" :  "4e3409e5-4e9a-4599-9b2e-f760b2fab380" ,  "ip_address" :  "192.168.0.11" }],  "port_id" :  "bbf99a69-0fb2-42a6-b7de-b7969bda9d73" ,  "net_id" :  "826be4f4-a23d-4c5c-bff5-7739936fac76" ,  "mac_addr" :  "fa:16:3e:01:04:2c" }]}

1	2016-10-17 12:06:30.078 29993 DEBUG octavia.controller.worker.tasks.amphora_driver_tasks [-] Finalized the amphora. execute  /usr/lib/python2 .7 /site-packages/octavia/controller/worker/tasks/amphora_driver_tasks .py:164

创建对外服务的vip的端口

1 2

2016-10-17 12:06:30.226 29993 DEBUG octavia.controller.worker.controller_worker [-] Task  'octavia.controller.worker.tasks.network_tasks.AllocateVIP'  (af8ea5a0-42c8-4d30-9ffa-016668811fc8) transitioned into state  'RUNNING'  from state  'PENDING'  _task_receiver  /usr/lib/python2 .7 /site-packages/taskflow/listeners/logging .py:189 2016-10-17 12:06:30.227 29993 DEBUG octavia.controller.worker.tasks.network_tasks [-] Allocate_vip port_id c7d7b552-83ac-4e0c-84bf-0b9cae661eab, subnet_id 74aea53a-014a-4f9c-86f9-805a2a772a27,ip_address 192.168.123.31 execute  /usr/lib/python2 .7 /site-packages/octavia/controller/worker/tasks/network_tasks .py:328

1	<br data - filtered = "filtered" >

在该vip下面创建一个实际的port 并把该port   attach至 云主机

1 2

2016-10-17 12:06:32.662 29993 DEBUG octavia.network.drivers.neutron.allowed_address_pairs [-] Created vip port: 1627d28d-bf54-46eb-9d78-410c5d647bf4  for  amphora: 3f6e22a1-e0b0-4098-ba20-daf47cfdae19 _plug_amphora_vip  /usr/lib/python2 .7 /site-packages/octavia/network/drivers/neutron/allowed_address_pairs .py:97 2016-10-17 12:06:32.663 29993 DEBUG novaclient.v2.client [-] REQ: curl -g -i -X POST http: //controller :8774 /v2 .1 /9a4b2de78c2d45cfbf6880dd34877f7b/servers/d3c97360-56b2-4f75-b905-2ef83870a342/os-interface  -H  "User-Agent: python-novaclient"  -H  "Content-Type: application/json"  -H  "Accept: application/json"  -H  "X-OpenStack-Nova-API-Version: 2.1"  -H  "X-Auth-Token: {SHA1}3f6ccac4cb8b70b06fb5e62b9db2272702d8ec67"  -d  '{"interfaceAttachment": {"port_id": "1627d28d-bf54-46eb-9d78-410c5d647bf4"}}'  _http_log_request  /usr/lib/python2 .7 /site-packages/keystoneauth1/session .py:337

创建listener

1 2 3

2016-10-17 19:01:09.384 29993 DEBUG octavia.amphorae.drivers.haproxy.rest_api_driver [-] request url https: //192 .168.0.9:9443 /0 .5 /listeners/c3a1867c-b2e5-49a7-819b-7a7d39063dda/reload  request  /usr/lib/python2 .7 /site-packages/octavia/amphorae/drivers/haproxy/rest_api_driver .py:248 2016-10-17 19:01:09.412 29993 DEBUG octavia.amphorae.drivers.haproxy.rest_api_driver [-] Connected to amphora. Response: <Response [202]> request  /usr/lib/python2 .7 /site-packages/octavia/amphorae/drivers/haproxy/rest_api_driver .py:270 2016-10-17 19:01:09.414 29993 DEBUG octavia.controller.worker.controller_worker [-] Task  'octavia.controller.worker.tasks.amphora_driver_tasks.ListenersUpdate'  (0f588287-a383-4c70-9847-20187dd19f9f) transitioned into state  'SUCCESS'  from state  'RUNNING'  with result  'None'  _task_receiver  /usr/lib/python2 .7 /site-packages/taskflow/listeners/logging .py:178

10.2

octavia agent分析

在9443端口创建监听端口给worker和health-manager 访问

1	2016-10-17 12:10:41.344 1043 INFO werkzeug [-]  * Running on https: //0 .0.0.0:9443/ (Press CTRL+C to quit)

octavia agent的似乎有bug，不显示debug信息。

11 高可用测试

将/etc/octavia/octavia.conf配置文件中的loadbalancer_topology = SINGLE 改成 ACTIVE_STANDBY 可以启用高可用模式，目前不持双ACTIVE

生成loadbalancer之后，可以看到生成两个虚拟机

[root@controller octavia]# neutron lbaas-loadbalancer-create --name test-lb1238 lbtest2 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Created a new loadbalancer: +---------------------+--------------------------------------+ | Field               | Value                                | +---------------------+--------------------------------------+ | admin_state_up      | True                                 | | description         |                                      | |  id                   | 4e43f3c7-c0f6-44c7-8dab-e2fc8ed16e0f | | listeners           |                                      | | name                |  test -lb1238                          | | operating_status    | OFFLINE                              | | pools               |                                      | | provider            | octavia                              | | provisioning_status | PENDING_CREATE                       | | tenant_id           | 9a4b2de78c2d45cfbf6880dd34877f7b     | | vip_address         | 192.168.235.14                       | | vip_port_id         | 42f72c9f-4623-4bf5-ae82-29f8cf588d2d | | vip_subnet_id       | 52e93565-eab2-4316-a04c-3e554992c993 | +---------------------+--------------------------------------+

1 2 3

[root@controller ~] # openstack server list |grep 192.168.235                 | | 736b8b76-2918-49a7-8477-995a168709bd | amphora-5379f109-01fa-429c-860b-c739e0c5ad5e | ACTIVE  | lb-mgmt-net=192.168.0.8; lbtest2=192.168.235.10  | amphora-x64-haproxy 201610131607    | | bd867667-b8d2-49c5-bb1e-54f0753d33a3 | amphora-23540889-b07e-4c0e-ab9b-df0075fbb9c3 | ACTIVE  | lb-mgmt-net=192.168.0.25; lbtest2=192.168.235.19 | amphora-x64-haproxy 201610131607

看到3个ip:vip是192.168.235.14,两台机器出口ip是192.168.235.10和192.168.235.19

登陆虚拟机验证一下，注意登陆的ip是管理网络的ip：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74

[root@controller ~] # ssh 192.168.0.8 "ps -ef |grep keepalived; cat  /var/lib/octavia/vrrp/octavia-keepalived.conf" root      1868     1  0 04:40 ?        00:00:00  /usr/sbin/keepalived  -D -d -f  /var/lib/octavia/vrrp/octavia-keepalived .conf root      1869  1868  0 04:40 ?        00:00:00  /usr/sbin/keepalived  -D -d -f  /var/lib/octavia/vrrp/octavia-keepalived .conf root      1870  1868  0 04:40 ?        00:00:00  /usr/sbin/keepalived  -D -d -f  /var/lib/octavia/vrrp/octavia-keepalived .conf root      5448  5377  0 05:00 ?        00:00:00  bash  -c  ps  -ef | grep  keepalived;  cat   /var/lib/octavia/vrrp/octavia-keepalived .conf root      5450  5448  0 05:00 ?        00:00:00  grep  keepalived vrrp_script check_script {    script  /var/lib/octavia/vrrp/check_script .sh    interval 5    fall 2    rise 2 } vrrp_instance 4e43f3c7c0f644c78dabe2fc8ed16e0f {   state MASTER   interface eth1   virtual_router_id 1   priority 100   nopreempt   garp_master_refresh 5   garp_master_refresh_repeat 2   advert_int 1   authentication {    auth_type PASS    auth_pass ee46125   }   unicast_src_ip 192.168.235.10   unicast_peer {         192.168.235.19   }   virtual_ipaddress {    192.168.235.14   }   track_script {      check_script   } } [root@controller ~] # ssh 192.168.0.8 "ps -ef |grep haproxy; cat  /var/lib/octavia/836053f0-ea72-46ae-9fae-8b80153ef593/haproxy.cfg" nobody    2195     1  0 04:43 ?        00:00:00  /usr/sbin/haproxy  -f  /var/lib/octavia/836053f0-ea72-46ae-9fae-8b80153ef593/haproxy .cfg -L jrwLnRhlvXcPd21JhvXEMStRHh0 -p  /var/lib/octavia/836053f0-ea72-46ae-9fae-8b80153ef593/836053f0-ea72-46ae-9fae-8b80153ef593 .pid -sf 2154 root      6745  6676  0 05:06 ?        00:00:00  bash  -c  ps  -ef | grep  haproxy;  cat   /var/lib/octavia/836053f0-ea72-46ae-9fae-8b80153ef593/haproxy .cfg root      6747  6745  0 05:06 ?        00:00:00  grep  haproxy # Configuration for test-lb1238 global      daemon      user nobody      group nogroup      log  /dev/log  local0      log  /dev/log  local1 notice      stats socket  /var/lib/octavia/836053f0-ea72-46ae-9fae-8b80153ef593 .sock mode 0666 level user   defaults      log global      retries 3      option redispatch      timeout connect 5000      timeout client 50000      timeout server 50000   peers 836053f0ea7246ae9fae8b80153ef593_peers      peer 3OduZJiPzm475Q7IgyshE5oq1Jk 192.168.235.19:1025      peer jrwLnRhlvXcPd21JhvXEMStRHh0 192.168.235.10:1025     frontend 836053f0-ea72-46ae-9fae-8b80153ef593      option tcplog      bind 192.168.235.14:22      mode tcp      default_backend 457d4de5-3213-4969-8f20-1f2d3505ff1e   backend 457d4de5-3213-4969-8f20-1f2d3505ff1e      mode tcp      balance leastconn      timeout check 5      server fa28676f-a762-4a8e-91ab-7a83f071b62b 192.168.235.20:22 weight 1 check inter 5s fall 3 rise 3      server 1ded44da-cba5-434c-8578-95153656c392 192.168.235.24:22 weight 1 check inter 5s fall 3 rise 3

另一台结果结果类似。

结论：octavia的高可用是通过haproxy加keepalived来完成的。

四、其他

1、在services_lbaas.conf下有个选项 

[octavia]

request_poll_timeout = 200

此选项的定义了，创建loadbalancer之后，当超过这个时间以后，如果octavia还没有的状态没有变成active，neutron就会把这个loadbalancer设置为error，默认值是100，在我的环境下高可用模式会来不及。日志如下：

1 2

2016-10-19 09:38:26.392 6256 DEBUG neutron_lbaas.drivers.octavia.driver [req-bee3619a-f9d4-4463-adcd-3cb99826b600 - - - - -] Octavia reports load balancer 2676dac6-c41d-4501-9c41-781a176c6baf has provisioning status of PENDING_CREATE thread_op  /usr/lib/python2 .7 /site-packages/neutron_lbaas/drivers/octavia/driver .py:75 2016-10-19 09:38:29.393 6256 DEBUG neutron_lbaas.drivers.octavia.driver [req-bee3619a-f9d4-4463-adcd-3cb99826b600 - - - - -] Timeout has expired  for  load balancer 2676dac6-c41d-4501-9c41-781a176c6baf to complete an operation.  The last reported status was PENDING_CREATE thread_op  /usr/lib/python2 .7 /site-packages/neutron_lbaas/drivers/octavia/driver .py:94

2、源代码小修改例子：

 当neutron的loadbalancer状态发生变成active或者error时候时候推送到报警系统

 修改/usr/lib/python2.7/site-packages/neutron_lbaas/drivers/octavia/driver.py

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

if  prov_status  = =  'ACTIVE'  or  prov_status  = =  'DELETED' :              kwargs  =  { 'delete' : delete}              if  manager.driver.allocates_vip  and  lb_create:                  kwargs[ 'lb_create' ]  =  lb_create                  # TODO(blogan): drop fk constraint on vip_port_id to ports                  # table because the port can't be removed unless the load                  # balancer has been deleted.  Until then we won't populate the                  # vip_port_id field.                  # entity.vip_port_id = octavia_lb.get('vip').get('port_id')                  entity.vip_address  =  octavia_lb.get( 'vip' ).get( 'ip_address' )              manager.successful_completion(context, entity,  * * kwargs)              if  prov_status  = =  'ACTIVE' :                urllib2.urlopen( 'http://********' )                LOG.debug( "report  status to******* {0}{1}" . format (entity.root_loadbalancer. id , prov_status))              return          elif  prov_status  = =  'ERROR' :              manager.failed_completion(context, entity)              urllib2.urlopen( 'http://*******' )              LOG.debug( "report status to ******* {0}{1}" . format (entity.root_loadbalancer. id , prov_status))              return

 

3、octavia的数据库和neutron不是同一张表，但是里面有很多数据要求要保持一致，一定要保持两者相关数据的同步，不同步的话会带来很多问题，亲身经历。

本文转自 superbigsea 51CTO博客，原文链接:http://blog.51cto.com/superbigsea/1862253