[cloud][OVS][sdn] Open vSwitch 初步了解

What is Open vSwitch?

Open vSwitch is a production quality, multilayer virtual switch licensed under the open source Apache 2.0 license. It is designed to enable massive network automation through programmatic extension, while still supporting standard management interfaces and protocols (e.g. NetFlow, sFlow, IPFIX, RSPAN, CLI, LACP, 802.1ag). In addition, it is designed to support distribution across multiple physical servers similar to VMware's vNetwork distributed vswitch or Cisco's Nexus 1000V. See full feature list here

Why Open vSwitch ?

https://github.com/openvswitch/ovs/blob/master/Documentation/intro/why-ovs.rst

OVN:

http://www.openvswitch.org/support/dist-docs/ovn-architecture.7.html

                                         CMS
                                          |
                                          |
                              +-----------|-----------+
                              |           |           |
                              |     OVN/CMS Plugin    |
                              |           |           |
                              |           |           |
                              |   OVN Northbound DB   |
                              |           |           |
                              |           |           |
                              |       ovn-northd      |
                              |           |           |
                              +-----------|-----------+
                                          |
                                          |
                                +-------------------+
                                | OVN Southbound DB |
                                +-------------------+
                                          |
                                          |
                       +------------------+------------------+
                       |                  |                  |
         HV           |                  |    HV n          |
       +---------------|---------------+  .  +---------------|---------------+
       |               |               |  .  |               |               |
       |        ovn-controller         |  .  |        ovn-controller         |
       |         |          |          |  .  |         |          |          |
       |         |          |          |     |         |          |          |
       |  ovs-vswitchd   ovsdb-server  |     |  ovs-vswitchd   ovsdb-server  |
       |                               |     |                               |
       +-------------------------------+     +-------------------------------+

在继续之前，做好先理解一下 namespace:

[cloud][sdn] network namespace

下面两篇中文介绍，内容相似，写的都不咋样。

http://fishcried.com/2016-02-09/openvswitch-ops-guide/

https://blog.kghost.info/2014/11/19/openvswitch-internal/

这个偏实践指导，写的好：

https://www.ibm.com/developerworks/cn/cloud/library/1401_zhaoyi_openswitch/

从源码编译：

文档：https://docs.openvswitch.org/en/latest/intro/install/general/

需要注意是否支持内核模块的编译，有所不同。

[root@D128 thirdparty]# git clone https://github.com/openvswitch/ovs.git
[root@D128 ovs]# git checkout v2.7.0
[root@D128 ovs]#  yum install autoconf automake libtool
[root@D128 ovs]# ./boot.sh
[root@D128 ovs]# ./configure --prefix=/root/BUILD_ovs/
[root@D128 ovs]# make
[root@D128 ovs]# make install

上边是没编译内核模块的。。。

再编个内核模块吧！！

[root@D128 ovs]# yum install kernel-devel-$(uname -r)
[root@D128 ovs]# ./configure --prefix=/root/BUILD_ovs/ --with-linux=/lib/modules/$(uname -r)/build
[root@D128 ovs]# uname -a
Linux D128 3.10.-.el7.x86_64 # SMP Tue Aug  :: UTC  x86_64 x86_64 x86_64 GNU/Linux

kernel版本太旧，编不过。降版本到v2.6.0

[root@D128 ovs]# git checkout v2.6.0
Previous HEAD position was c298ef7... Set release date for 2.7..
HEAD is now at 7a0f907... Set release date for 2.6..
[root@D128 ovs]# git branch
* (detached from v2.6.0)
  master
[root@D128 ovs]#

还是编不过。如果想编译通过必须要找对彼此兼容的版本。算了直接用centos自带的ko（反正是学习了解阶段，但愿能兼容跑起来）。。。

[root@D128 ovs]# modprobe openvswitch
[root@D128 ovs]# lsmod|grep openvswitch
openvswitch
nf_nat_ipv6               openvswitch
nf_nat_ipv4               openvswitch
nf_defrag_ipv6            openvswitch,nf_conntrack_ipv6
nf_nat                    openvswitch,nf_nat_ipv4,nf_nat_ipv6
nf_conntrack             openvswitch,nf_nat,nf_nat_ipv4,nf_nat_ipv6,nf_conntrack_ipv4,nf_conntrack_ipv6
libcrc32c                 xfs,openvswitch,nf_nat,nf_conntrack
[root@D128 ovs]#

运行：

[root@D128 ovs]# export PATH=$PATH:/root/BUILD_ovs/share/openvswitch/scripts/
[root@D128 ~]# ovs-ctl --system-id=random start
Starting ovsdb-server                                      [  OK  ]
Configuring Open vSwitch system IDs                        [  OK  ]
Starting ovs-vswitchd                                      [  OK  ]
Enabling remote OVSDB managers                             [  OK  ]
[root@D128 ~]#

数据库应该已经建立了默认的，并且都初始化了。

测试：

[root@D128 BUILD_ovs]# ip link
: lo: <LOOPBACK,UP,LOWER_UP> mtu  qdisc noqueue state UNKNOWN mode DEFAULT qlen
    link/loopback ::::: brd :::::
: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP mode DEFAULT qlen
    link/ether :0c::2f:cf: brd ff:ff:ff:ff:ff:ff
[root@D128 BUILD_ovs]# ./bin/ovs-vsctl add-br br0
[root@D128 BUILD_ovs]# ip link
: lo: <LOOPBACK,UP,LOWER_UP> mtu  qdisc noqueue state UNKNOWN mode DEFAULT qlen
    link/loopback ::::: brd :::::
: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP mode DEFAULT qlen
    link/ether :0c::2f:cf: brd ff:ff:ff:ff:ff:ff
3: ovs-system: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether 92:e5:c6:d2:ec:a2 brd ff:ff:ff:ff:ff:ff
4: br0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT qlen 1000
    link/ether be:e8:bd:df:ff:41 brd ff:ff:ff:ff:ff:ff
[root@D128 BUILD_ovs]# ./bin/ovs-vsctl add-port br0 ens33
[root@D128 BUILD_ovs]# brctl show
bridge name    bridge id        STP enabled    interfaces
[root@D128 BUILD_ovs]#

ovs-vsctl add-br 增加的这俩个设备是什么？

我写了个脚本去判断：

[root@D128 ~]# cat ip_link_show_type.sh
#! /bin/bash
 
TYPE=" vlan veth vcan dummy ifb macvlan macvtap bridge bond ipoib ip6tnl ipip sit vxlan gre gretap ip6gre ip6gretap vti nlmon bond_slave geneve bridge_slave macsec"
 
for T in $TYPE
do
    echo $T
    ip link show type $T
done

竟然不属于这里边任意一个类型。。。。

虽然没有正面回答，但是看一下下面这个问题的答案，就能很好的理解了：

同时，也讲解了KVM的tap设备怎么可OVS协作，以及需要注意什么。

https://github.com/openvswitch/ovs/blob/master/Documentation/faq/issues.rst

Q: I created a tap device tap0, configured an IP address on it, and added it to a bridge, like this:

大概就是说，OVS的bridge和OVS的internal port是OVS单独实现的两种特殊设备。

报错：

[root@D128 BUILD_ovs]# ./bin/ovs-vsctl add-port br0 p0
ovs-vsctl: Error detected while setting up 'p0': could not open network device p0 (No such device).  See ovs-vswitchd log for details.
ovs-vsctl: The default log directory is "/root/BUILD_ovs/var/log/openvswitch".
[root@D128 BUILD_ovs]#

要这样：

https://github.com/openvswitch/ovs-issues/issues/110

The port‘s name should be a exist interface use ifconfig to see, such as eth0. If you just want to use a virtual port name to make a test you should 
specify the port's type like ovs-vsctl add-port br0 port0 -- set Interface port0 type=internal or ovs-vsctl set Interface port0 type=internal

[root@D128 BUILD_ovs]# ./bin/ovs-vsctl add-port br0 port0 -- set Interface port0 type=internal
[root@D128 BUILD_ovs]# ip link
: lo: <LOOPBACK,UP,LOWER_UP> mtu  qdisc noqueue state UNKNOWN mode DEFAULT qlen
    link/loopback ::::: brd :::::
: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu  qdisc pfifo_fast state UP mode DEFAULT qlen
    link/ether :0c::2f:cf: brd ff:ff:ff:ff:ff:ff
: ovs-system: <BROADCAST,MULTICAST> mtu  qdisc noop state DOWN mode DEFAULT qlen
    link/ether :e5:c6:d2:ec:a2 brd ff:ff:ff:ff:ff:ff
: br0: <BROADCAST,MULTICAST> mtu  qdisc noop state DOWN mode DEFAULT qlen
    link/ether be:e8:bd:df:ff: brd ff:ff:ff:ff:ff:ff
: port0: <BROADCAST,MULTICAST> mtu  qdisc noop state DOWN mode DEFAULT qlen
    link/ether f6:cb:c2::fc:e0 brd ff:ff:ff:ff:ff:ff
[root@D128 BUILD_ovs]# ./bin/ovs-vsctl show
528b5679-22e8-484b-947b-4499959dc341
    Bridge "br0"
        Port "port0"
            Interface "port0"
                type: internal
        Port "br0"
            Interface "br0"
                type: internal
    ovs_version: "2.7.0"
[root@D128 BUILD_ovs]#

查看br0，port0这两个设备。

[root@D128 BUILD_ovs]# ethtool -i br0
driver: openvswitch
version:
firmware-version:
expansion-rom-version:
bus-info:
supports-statistics: no
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no
[root@D128 BUILD_ovs]# ethtool -i port0
driver: openvswitch
version:
firmware-version:
expansion-rom-version:
bus-info:
supports-statistics: no
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no
[root@D128 BUILD_ovs]#

增加namespace

[root@D128 BUILD_ovs]# ip netns add ns0
[root@D128 BUILD_ovs]# ip link set port0 netns ns0

[root@D128 BUILD_ovs]# ip netns exec ns0 ip addr add 192.168.1.100/ dev port0
[root@D128 BUILD_ovs]# ip netns exec ns0 ifconfig port0 promisc up

查看：

[root@D128 BUILD_ovs]# ./bin/ovs-ofctl show br0
OFPT_FEATURES_REPLY (xid=0x2): dpid:0000bee8bddfff41
n_tables:, n_buffers:
capabilities: FLOW_STATS TABLE_STATS PORT_STATS QUEUE_STATS ARP_MATCH_IP
actions: output enqueue set_vlan_vid set_vlan_pcp strip_vlan mod_dl_src mod_dl_dst mod_nw_src mod_nw_dst mod_nw_tos mod_tp_src mod_tp_dst
 (port0): addr:::::0c:e1
     config:     PORT_DOWN
     state:      LINK_DOWN
     speed:  Mbps now,  Mbps max
 LOCAL(br0): addr:be:e8:bd:df:ff:
     config:     PORT_DOWN
     state:      LINK_DOWN
     speed:  Mbps now,  Mbps max
OFPT_GET_CONFIG_REPLY (xid=0x4): frags=normal miss_send_len=
[root@D128 BUILD_ovs]# ./bin/ovs-dpctl show
system@ovs-system:
    lookups: hit: missed: lost:
    flows:
    masks: hit: total: hit/pkt:1.21
    port : ovs-system (internal)
    port : br0 (internal)
    port : port0 (internal)
[root@D128 BUILD_ovs]#

[root@D128 BUILD_ovs]# ip addr add 192.168.1.101/ dev br0
[root@D128 BUILD_ovs]# ip link set br0 up

现在两个namespace可以通过br0 互通了。

增加一个of规则

[root@D128 BUILD_ovs]# ./bin/ovs-ofctl add-flow br0 "priority=1 idle_timeout=0, in_port=2,actions=mod_nw_src:9.181.137.1,normal"
[root@D128 BUILD_ovs]# ./bin/ovs-ofctl show br0
OFPT_FEATURES_REPLY (xid=0x2): dpid:0000bee8bddfff41
n_tables:, n_buffers:
capabilities: FLOW_STATS TABLE_STATS PORT_STATS QUEUE_STATS ARP_MATCH_IP
actions: output enqueue set_vlan_vid set_vlan_pcp strip_vlan mod_dl_src mod_dl_dst mod_nw_src mod_nw_dst mod_nw_tos mod_tp_src mod_tp_dst
 (port0): addr:::::0c:e1
     config:     PORT_DOWN
     state:      LINK_DOWN
     speed:  Mbps now,  Mbps max
 LOCAL(br0): addr:be:e8:bd:df:ff:
     config:
     state:
     speed:  Mbps now,  Mbps max
OFPT_GET_CONFIG_REPLY (xid=0x4): frags=normal miss_send_len=
[root@D128 BUILD_ovs]# ./bin/ovs-

抓包可以看见，原地址已经被修改为9.181.137.1

[root@D128 BUILD_ovs]# tcpdump -i br0
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on br0, link-type EN10MB (Ethernet), capture size  bytes
::00.988146 IP 9.181.137.1 > localhost: ICMP echo request, id , seq , length
::01.988227 IP 9.181.137.1 > localhost: ICMP echo request, id , seq , length
::02.988113 IP 9.181.137.1 > localhost: ICMP echo request, id , seq , length
::03.988133 IP 9.181.137.1 > localhost: ICMP echo request, id , seq , length

暂时先到这。作为一个初步了解。

-------------------------------------------------- update @ 2018-03-30 20：11 -----------------------------------

在OVS port上抓包。

把port流量镜像出来。

[root@vrouter-ovs ~]# ip link add dev mirror type dummy
[root@vrouter-ovs ~]# ip link set mirror up
 
[root@dr-lb ~]# ovs-vsctl add-port ovs-br0 mirror

[root@vrouter-ovs ~]# ovs-vsctl -- set Bridge br-tun mirror=@mi -- --id=@pmirror get Port mirror -- --id=@patch get Port tun-to-int -- --id=@mi create Mirror name=mymi select-dst-port=@patch select-src-port=@patch output-port=@pmirror
[root@vrouter-ovs ~]# tcpdump -i mirror -nn

查看和删除mirror

# ovs-vsctl list Mirror
# ovs-vsctl clear bridge ovsbr0 mirrors

ovs-vsctl set Bridge ovs-br0 mirrors=@mi -- --id=@pmirror get Port mirror-br0 -- --id=@patch get port vxlanclient0 -- --id=@mi create Mirror name=mymi select-dst-port=@patch select-src-port=@patch output-port=@pmirror