【OpenStack】OpenStack系列11之namaspace&openvswitch原理实践

  1. Namespace实现网络隔离与互通
    1. 新建ns:

      ip netns add foo

    2. 查看ns:

      ip netns

    3. 查看ns详细配置:

      ip netns exec foo ip addr

    4. 设置ns内部lo网络up:

      ip netns exec foo ip link set lo up

      ip netns exec foo ip addr

      ip netns exec foo route –n

      ip netns exec foo iptables -t nat –S

    5. ip netns exec foo ping 10.20.0.88

    6. 创建tap设备对:

      ip link add tap-foo type veth peer name tap-root

      ip link

      使用ethtool确认属于成对设备:

    7. 设置tap设备的ns:

      ip link set tap-foo netns foo

      ip netns exec foo ip link

      ip netns exec foo ip link set tap-foo up

      root的ns看不到foo-tap了:

    8. 为tap设备添加ip地址

      ip netns exec foo ip addr add 88.88.88.100/24 dev tap-foo

      ip netns exec foo ifconfig

    9. 设置tap-root设备的ip地址

      ip addr add 88.88.88.101/24 dev tap-root

    10. up起root的tap设备

      ip link set tap-root up

      ifconfig

    11. 可以相互ping通88.88.88.100、88.88.88.101,但是仍然不可访问10.20.0.88

      ip netns exec foo ping 10.20.0.88

    12. 设置默认路由

      ip netns exec foo route add -net 0.0.0.0 netmask 0.0.0.0 gw 88.88.88.101

      ip netns exec foo route –n

    13. 继续ping10.20.0.88,可以ping通

      ip netns exec foo ping 10.20.0.88

    14. 机器重启所有配置自动清除。
  2. Openvswitch VM+GRE隧道实现同网段虚拟机跨主机通信
    1. 环境拓扑

    2. 环境准备

      network1

    eth0:10.20.0.201 (management network)

    eht1:172.16.0.201 (public/external network)

    eht2:192.168.4.201 (private network,gre tunning)

     

    compute1

    eth0:10.20.0.202 (management network)

    eht1:(disabled)

    eht2:192.168.4.202 (private network,gre tunning)

    1. 先看网络节点。

      安装必备软件包:yum install libvirt openvswitch python-virtinst xauth tigervnc –y

    yum install kvm qemu-kvm qemu-kvm-tools

    yum install bridge-utils qemu-img

    yum install python-virtinst virt-manager virt-viewer

    yum install libvirt libvirt-client

    reboot -h 0,vm重启生效。

    1. 移除默认的libvirt网络

    virsh net-destroy default

    virsh net-autostart --disable default

    virsh net-undefine default

    virsh net-list

    1. 设置允许ipforward

    vi /etc/sysctl.conf

    net.ipv4.ip_forward=1

    net.ipv4.conf.all.rp_filter=0

    net.ipv4.conf.default.rp_filter=0

    sysctl –p,立即生效。

    1. 启动openvswitch

      service openvswitch start

      chkconfig openvswitch on

    2. 创建linux bridge,qbr设备,虚拟机网卡与br-int网桥之间的桥梁网桥。

      brctl addbr qbr01

      ip link set qbr01 up

      brctl show

      ifconfig

    3. 创建虚拟机,虚拟机网卡连接到网桥qbr01的tap设备:

      mkdir -p /var/tmp/gre

      mv instance1.img instance1.xml /var/tmp/gre

      instance文件参考:

      https://github.com/yongluo2013/osf-openstack-training/blob/master/installation/gre/instance1.xml

      cd /var/tmp/gre

      virsh define instance1.xml

      virsh start instance1

      virsh vncdisplay instance1

      vncviewer :0

      vncviewer进入虚拟机。

    4. 为虚拟机配置IP与默认网关

      sudo ip addr add 192.168.1.11/24 dev eth0

      sudo route add default gw 192.168.1.1

    5. 创建br-int网桥,用于连接qbr网桥

      ovs-vsctl add-br br-int

      ovs-vsctl add-port br-int gre0 -- set interface gre0 type=gre options:remote_ip=192.168.4.202

    6. 创建tap对,用于br-int与qbr之间的连接

      ip link add qvo01 type veth peer name qvb01

      brctl addif qbr01 qvb01,设置qvb01为qbr01的接口

      ovs-vsctl add-port br-int qvo01

      ovs-vsctl set port qvo01 tag=100

      ip link set qvb01 up

      ip link set qvo01 up

      br-int以下的网络设备均up起来了

    7. 再看计算节点。准备软件与网络节点一致。

      brctl addbr qbr02

      ip link set qbr02 up

       

      virsh define instance2.xml

      virsh start instance2

      virsh vncdisplay instance2

      vncviewer :0

       

      sudo ip addr add 192.168.1.12/24 dev eth0

      sudo route add default gw 192.168.1.1

       

      ovs-vsctl add-br br-int

      ovs-vsctl add-port br-int gre0 -- set interface gre0 type=gre options:remote_ip=192.168.4.201

       

      ip link add qvo02 type veth peer name qvb02

      brctl addif qbr02 qvb02

      ovs-vsctl add-port br-int qvo02

      ovs-vsctl set port qvo02 tag=100

      ip link set qvb02 up

      ip link set qvo02 up

       

    8. 至此,vm01的192.168.1.11可以ping通vm02的192.168.1.12。在没有namespace场景下,同一vlan可以直接连通。

      vm01:

       

      vm02:

  3. Openvswitch VM+GRE隧道+NameSpace实现dhcp的namespace
    1. 添加dhcp的ns

      ip netns add dhcp01

    2. 为br-int添加dhcp的tap设备接口

      ovs-vsctl add-port br-int tapdhcp01 -- set interface tapdhcp01 type=internal

    3. ovs-vsctl set port tapdhcp01 tag=100

    4. 设置tap设备所属的ns

      ip link set tapdhcp01 netns dhcp01

    5. 为dhcp的tap设备设置ip

      ip netns exec dhcp01 ip addr add 192.168.1.2/24 dev tapdhcp01

      ip netns exec dhcp01 ip link set tapdhcp01 up

    6. 在namespace中可以访问vm01与vm02。
  4. Openvswitch VM+GRE隧道+NameSpace实现router的namespace
    1. ovs-vsctl add-br br-ex
    2. 重新配置eth1和br-ex

    vi /etc/sysconfig/network-scripts/ifcfg-eth1

    DEVICE=eth1

    ONBOOT=yes

    BOOTPROTO=none

    PROMISC=yes

    MTU=1546

     

    vi /etc/sysconfig/network-scripts/ifcfg-br-ex

    DEVICE=br-ex

    TYPE=Bridge

    ONBOOT=yes

    BOOTPROTO=none

    IPADDR0=172.16.0.201

    PREFIX0=24

     

    ovs-vsctl add-port br-ex eth1

    1. 重启网络服务:service network restart

      ip link set br-ex up

      sudo ip addr add 172.16.0.20/24 dev br-ex

    2. 添加router的ns

      ip netns add router01

    3. 添加私网的网关

      ovs-vsctl add-port br-int qr01 -- set interface qr01 type=internal

      ovs-vsctl set port qr01 tag=100

    4. 设置网关在router01的ns中

      ip link set qr01 netns router01

      ip netns exec router01 ip addr add 192.168.1.1/24 dev qr01

      ip netns exec router01 ip link set qr01 up

      ip netns exec router01 ip link set lo up

    5. 至此,路由器的qr端可以ping通vm的私有ip。

    6. 在br-ex中添加qg设备,设置为私有网段连接外网的下一跳地址

    ovs-vsctl add-port br-ex qg01 -- set interface qg01 type=internal

    ip link set qg01 netns router01

    ip netns exec router01 ip addr add 172.16.0.100/24 dev qg01

    ip netns exec router01 ip link set qg01 up

    ip netns exec router01 ip link set lo up

     

    1. 为虚拟机添加浮动ip,并且配置iptables,规则使虚拟机可以访问外网

      ip netns exec router01 ip addr add 172.16.0.101/32 dev qg01

      ip netns exec router01 iptables -t nat -A OUTPUT -d 172.16.0.101/32 -j DNAT --to-destination 192.168.1.11

      ip netns exec router01 iptables -t nat -A PREROUTING -d 172.16.0.101/32 -j DNAT --to-destination 192.168.1.11

      ip netns exec router01 iptables -t nat -A POSTROUTING -s 192.168.1.11/32 -j SNAT --to-source 172.16.0.101

      ip netns exec router01 iptables -t nat -A POSTROUTING -s 192.168.1.0/24 -j SNAT --to-source 172.16.0.100

    2. 到此,可以使用洗浮动IP访问虚拟机,如下图所示:

  5. 参考文档:

    https://github.com/yongluo2013/osf-openstack-training/blob/master/installation/mock-neutron-sdn-with-ovs-gre.md

posted @ 2016-03-09 11:02  junneyang  阅读(1880)  评论(0编辑  收藏  举报