分享

Neutron配置文件简析(J版RDO安装)

w123aw 发表于 2015-11-23 17:46:06 [显示全部楼层] 回帖奖励 阅读模式 关闭右栏 1 12123


Neutron的配置文件随着代码的修改而变化,理解每个配置选项的含义对于正确部署Neutron或快速定位Neutron的异常问题有着重要的意思。下面仅仅简陋的讲解下Neutron主要配置文件重要配置项的含义,表述不准确的地方还请见谅。



1)neutron进程分布

控制节点:neutron-server,

网络节点:

neutron-lbaas-agent/neutron-metadata-agent/neutron-l3-agent/neutron-dhcp-agent/neutron-openvswitch-agent/neutron-vpn-agent(可以看出没有fwaas的服务进程,原因是其包含在neutron-l3-agent中)

计算节点:neutron-openvswitch-agent;

每个进程所用到的配置文件,可以通过ps –aux|grep  service_name来查看启动所需的相应文件。

2)配置文件主要有:

其中J版中新添dnsmasq-neutron.conf以增加对dnsmasq服务的控制,下面对配置文件的主要参数进行解析:

a)cat api-paste.ini

用于neutron提供restful api被调用的认证相关所需信息,包括租户名称/用户名称/密码、主机IP/访问协议http/访问端口35357等token信息:

b)cat dnsmasq-neutron.conf

dhcp-option-force=26,1400

其中1400代表新建VM通过DCHP获取IP后端口的MTU大小;

c)cat l3_agent.ini

[DEFAULT]

#L3 agent服务调试开关

debug = False

# OVS, Ryu, NEC做为Plugin时的接口驱动

interface_driver=neutron.agent.linux.interface.OVSInterfaceDriver

#OVS interface 是否使用ovs_use_veth,需要内核支持namespace.

# ovs_use_veth = False



# 允许建立多个namespace,以此来支持IP地址重叠,即一个租户在不同的路由器下可以同时使用相同的私网网段

use_namespaces = True



# 此选项在内核不支持多个namespace时,只绑定一个对应的router

router_id =



# 在创建多个外网时(不包含一个外网多个子网的情况),用于将l3 agent和外网ID的绑定

gateway_external_network_id =



# Indicates that this L3 agent should alsohandle routers that do not have

# an external network gatewayconfigured.  This option should be Trueonly

# for a single agent in a Neutrondeployment, and may be False for all agents

# if all routers must have an externalnetwork gateway

# handle_internal_only_routers = True

handle_internal_only_routers = True



# 网络节点外网网桥的名字

external_network_bridge = br-ex

# Neutron metadata server的TCP端口

metadata_port = 9697



# 同步路由相关的配置项

send_arp_for_ha = 3

periodic_interval = 40

periodic_fuzzy_delay = 5

# 使能metadata_proxy

enable_metadata_proxy = True

# Location of Metadata Proxy UNIX domainsocket

# metadata_proxy_socket =$state_path/metadata_proxy

# 默认情况下,当虚拟路由器删除时,对应的namespace是不会删除的,这样做的原因可以详见https://bugs.launchpad.net/neutron/+bug/1052535的讨论;开启时需明确风险

router_delete_namespaces = False

# 执行ovs命令的超时时间

# ovs_vsctl_timeout = 10

# 该选项是针对J版新添的DVR特性做的配置项,默认legacy是没有开启DVR的时候配置,开启DVR的时候,计算节点配置为dvr,单纯的网络节点配置为dvr_snat

agent_mode = legacy



# l3 agnet ha的一些配置选项

# ha_confs_path = $state_path/ha_confs

# ha_vrrp_auth_type = PASS

# ha_vrrp_auth_password =

# ha_vrrp_advert_int = 2

#是否自动failover

allow_automatic_l3agent_failover=False



d)metadata_agent.ini文件主要用于获取。。的信息,和网络本身关系不大;

e)plugin.ini为plugins/ml2/ml2_conf.ini的链接,仅在控制节点有,其内容解析参见ml2_conf.in;

f) policy.json为控制策略相关的配置,如果/etc/selinux/config为SELINUX=disabled或SELINUX=permissive将不再生效,对于安全要求较高的平台不能关闭或失效该服务;

g)rootwrap.conf的主要作用是让neutron的python代码配置,最终都要转换为底层core plugin或service plugin的在下发命令,比如linux namespace、ovs、iptables、haproxy、openswan等等,在配置这些底层服务命令的时候,需要保证命令的可执行权限或合适路径;

h)cat dhcp_agent.ini

DHCP配置选项中和多和l3 agent重复,其中需要提及的包括租约个数以防止攻击,以及与router的namespace处理方式相对应的namspace的处理配置选项;

………

# Limit number of leases to prevent adenial-of-service.

# dnsmasq_lease_max = 16777216

……….

# dhcp_delete_namespaces, which is false bydefault, can be set to True if

# namespaces can be deleted cleanly on thehost running the dhcp agent.

# Do not enable this until you understandthe problem with the Linux iproute

# utility mentioned inhttps://bugs.launchpad.net/neutron/+bug/1052535 and

# you are sure that your version of iproutedoes not suffer from the problem.

# If True, namespaces will be deleted whena dhcp server is disabled.

# dhcp_delete_namespaces = False

dhcp_delete_namespaces = False

…………..

i) catfwaas_driver.ini

#比较简单,主要是底层驱动以及是否使能,fwaas的底层plugin默认为iptables

[fwaas]

driver =neutron.services.firewall.drivers.linux.iptables_fwaas.IptablesFwaasDriver

enabled = True

j) cat lbaas_agent.ini

配置选项较多,较为重要的和fwaas一样,底层的plugin配置

…..

device_driver =neutron.services.loadbalancer.drivers.haproxy.namespace_driver.HaproxyNSDriver

[haproxy]

……

# The user group

user_group = haproxy



k) cat neutron.conf

#neutron的配置选项较多,仅对重要选项进行一一解释

[DEFAULT]

……..

# 是否开启DVR功能,如果开启需要底层做相关的配置;默认为False表示不开启

router_distributed = False

……..

#日志目录

log_dir =/var/log/neutron

………….

#提供访问API的地址和端口

bind_host = 0.0.0.0

bind_port = 9696



………..

# core_plugin 的配置,J版后ml2用途较多

core_plugin =neutron.plugins.ml2.plugin.Ml2Plugin

#需要开启的services功能,可能包括router,firewall,lbaas,vpnaas,metering

service_plugins=neutron.services.loadbalancer.plugin.LoadBalancerPlugin,neutron.services.l3_router.l3_router_plugin.L3RouterPlugin,neutron.services.firewall.fwaas_plugin.FirewallPlugin

………..

#虚拟端口的分配基址

base_mac = fa:16:3e:00:00:00



# 相应的DVR Base MACaddress

# dvr_base_mac = fa:16:3f:00:00:00

…..

#DHCP的租约时间

dhcp_lease_duration = 86400

…………….

# 是否允许地址重叠,如果允许的话底层需要支持namespace功能

allow_overlapping_ips = True

…………..

对于网络配置有关限额的配置

# Maximum number of DNS nameservers persubnet

# max_dns_nameservers = 5

# Maximum number of host routes per subnet

# max_subnet_host_routes = 20

# Maximum number of fixed ips per port

# max_fixed_ips_per_port = 5

# Maximum number of routes per router

# max_routes = 30



# 用来判断当neutron的agent最后更新时间和当前时间(neutronserver)相差超过该值时,就认为该agent是死掉的

agent_down_time = 75



#当有多个l3 agent时,路由,dhcp,lb等的调度策略,默认都是ChanceScheduler

# network_scheduler_driver =neutron.scheduler.dhcp_agent_scheduler.ChanceScheduler

router_scheduler_driver =neutron.scheduler.l3_agent_scheduler.ChanceScheduler

# loadbalancer_pool_scheduler_driver =neutron.services.loadbalancer.agent_scheduler.ChanceScheduler

# network_auto_schedule = True

# router_auto_schedule = True



# 自动将死掉的l3 agent重新分配到其他活着的l3agent上,需要admin_state_up被设置为alive.默认为False表示关闭该功能

# allow_automatic_l3agent_failover = False



# 对应一个网络开启的dhcp服务数,当有多个l3agent时会分布在不同的l3 agent上起到HA的作用

dhcp_agents_per_network = 3



# 是否开启L3 HA的作用以及开启的相关配置选项

# l3_ha = False

## max_l3_agents_per_router = 3

# min_l3_agents_per_router = 2

# l3_ha_net_cidr = 169.254.192.0/18

………….

#为提高api和rpc的服务能力,可以开启多个服务进程,I版默认为0,表示值开启一个,J版的RDO上默认开启16个

api_workers = 2

rpc_workers = 2



#当端口改变时,通知nova

notify_nova_on_port_status_changes = True

notify_nova_on_port_data_changes = True

nova_url = http://192.168.23.235:8774/v2

nova_region_name =RegionOne

nova_admin_username =nova

nova_admin_tenant_id =42736627a3c1418aa76b0f6f5db3c3e1

nova_admin_password =9cce0a69ebfa4e8d

nova_admin_auth_url =http://192.168.23.235:35357/v2.0



#和其他组件一样,可以选择对应的消息机制类型,包括amqp,qpid等,也包括rabbit的消息队列集群HA服务

………….

#amqp_durable_queues=false

…………..

#qpid_hostname=localhost

…………

rabbit_port=5672

rabbit_hosts=192.168.23.235:5672



# ZeroMQ receiver listening port. (integervalue)

#rpc_zmq_port=9501

……………..



[quotas]

# 网络相关限额配置

# default_quota = -1

# quota_network = 10

# quota_subnet = 10

# quota_port = 50

# quota_security_group = 10

# quota_security_group_rule = 100

# quota_vip = 10

# quota_pool = 10

# quota_member = -1

# quota_health_monitor = -1

# quota_router = 10

# quota_floatingip = 50

# quota_firewall = 1

# quota_firewall_policy = 1

# quota_firewall_rule = 100



[agent]

#agent想server报告状态的时间间隔,

report_interval = 30

…………

[service_providers]

# 向neutron提供底层serviceplugin的说明

service_provider=LOADBALANCER:Haproxy:neutron.services.loadbalancer.drivers.haproxy.plugin_driver.HaproxyOnHostPluginDriver:default

service_provider=VPN:openswan:neutron.services.vpn.service_drivers.ipsec.IPsecVPNDriver:default



k)cat plugins/ml2/ml2_conf.ini

[ml2]

# 底层网络资源使用类型type_drivers= local,flat,vlan,gre,vxlan

type_drivers = vxlan

# Example: type_drivers =flat,vlan,gre,vxlan

# 租户网络类型,可以配置多个tenant_network_types= vlan,gre,vxlan

tenant_network_types = vxlan

# 底层驱动机制

mechanism_drivers =openvswitch



[ml2_type_flat]

…………..

#flat对应物理网络配置:

# Example:flat_networks = physnet1,physnet2



[ml2_type_vlan]

…………..

#vlan对应的物理网络配置,包括vlan的范围

# network_vlan_ranges =

# Example: network_vlan_ranges =physnet1:1000:2999,physnet2



[ml2_type_gre]

# gre类型对应的隧道ID范围

# tunnel_id_ranges =



[ml2_type_vxlan]

# vxlan对应的vni范围

vni_ranges =1:1000

#vxlan对应的组播组

vxlan_group =224.0.0.1



[securitygroup]

#使能安全组

enable_security_group = True

………….

l) catplugins/openvswitch/ovs_neutron_plugin.ini

[ovs]

…………….

#ovs对应的底层配置选项,需要和上面ml2的机制相对应,包括网络类型以及对应的取值范围,参与含义根据名字很容易解析出,不再赘述

# Sample Configurations.

#-----------------------------------------------------------------------------

#

# 1. With VLANs on eth1.

# [ovs]

# network_vlan_ranges = default:2000:3999

# tunnel_id_ranges =

# integration_bridge = br-int

# bridge_mappings = default:br-eth1

#

# 2. With GRE tunneling.

# [ovs]

# network_vlan_ranges =

# tunnel_id_ranges = 1:1000

# integration_bridge = br-int

# tunnel_bridge = br-tun

# local_ip = 10.0.0.3

#

# 3. With VXLAN tunneling.

# [ovs]

# network_vlan_ranges =

# tenant_network_type = vxlan

# tunnel_type = vxlan

# tunnel_id_ranges = 1:1000

# integration_bridge = br-int

# tunnel_bridge = br-tun

# local_ip = 10.0.0.3

# [agent]

# tunnel_types = vxlan

n) cat vpn_agent.ini

[DEFAULT]

…….

#与FW和LB一样,重要的是底层driver

[vpnagent]

…………..

vpn_device_driver=neutron.services.vpn.device_drivers.ipsec.OpenSwanDriver

[ipsec]

# ipsec_status_check_interval=60

        sina微博@北京-小武  个人博客: http://blog.csdn.net/night_elf_1020

已有(1)人评论

跳转到指定楼层
lmlm1234 发表于 2015-11-24 10:34:47
neutron进程分布
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

关闭

推荐上一条 /2 下一条