1 Open vSwitch <http://openvswitch.org>
3 Frequently Asked Questions
4 ==========================
9 Q: What is Open vSwitch?
11 A: Open vSwitch is a production quality open source software switch
12 designed to be used as a vswitch in virtualized server
13 environments. A vswitch forwards traffic between different VMs on
14 the same physical host and also forwards traffic between VMs and
15 the physical network. Open vSwitch supports standard management
16 interfaces (e.g. sFlow, NetFlow, IPFIX, RSPAN, CLI), and is open to
17 programmatic extension and control using OpenFlow and the OVSDB
20 Open vSwitch as designed to be compatible with modern switching
21 chipsets. This means that it can be ported to existing high-fanout
22 switches allowing the same flexible control of the physical
23 infrastructure as the virtual infrastructure. It also means that
24 Open vSwitch will be able to take advantage of on-NIC switching
25 chipsets as their functionality matures.
27 Q: What virtualization platforms can use Open vSwitch?
29 A: Open vSwitch can currently run on any Linux-based virtualization
30 platform (kernel 2.6.32 and newer), including: KVM, VirtualBox, Xen,
31 Xen Cloud Platform, XenServer. As of Linux 3.3 it is part of the
32 mainline kernel. The bulk of the code is written in platform-
33 independent C and is easily ported to other environments. We welcome
34 inquires about integrating Open vSwitch with other virtualization
37 Q: How can I try Open vSwitch?
39 A: The Open vSwitch source code can be built on a Linux system. You can
40 build and experiment with Open vSwitch on any Linux machine.
41 Packages for various Linux distributions are available on many
42 platforms, including: Debian, Ubuntu, Fedora.
44 You may also download and run a virtualization platform that already
45 has Open vSwitch integrated. For example, download a recent ISO for
46 XenServer or Xen Cloud Platform. Be aware that the version
47 integrated with a particular platform may not be the most recent Open
50 Q: Does Open vSwitch only work on Linux?
52 A: No, Open vSwitch has been ported to a number of different operating
53 systems and hardware platforms. Most of the development work occurs
54 on Linux, but the code should be portable to any POSIX system. We've
55 seen Open vSwitch ported to a number of different platforms,
56 including FreeBSD, Windows, and even non-POSIX embedded systems.
58 By definition, the Open vSwitch Linux kernel module only works on
59 Linux and will provide the highest performance. However, a userspace
60 datapath is available that should be very portable.
62 Q: What's involved with porting Open vSwitch to a new platform or
65 A: The PORTING document describes how one would go about porting Open
66 vSwitch to a new operating system or hardware platform.
68 Q: Why would I use Open vSwitch instead of the Linux bridge?
70 A: Open vSwitch is specially designed to make it easier to manage VM
71 network configuration and monitor state spread across many physical
72 hosts in dynamic virtualized environments. Please see WHY-OVS for a
73 more detailed description of how Open vSwitch relates to the Linux
76 Q: How is Open vSwitch related to distributed virtual switches like the
77 VMware vNetwork distributed switch or the Cisco Nexus 1000V?
79 A: Distributed vswitch applications (e.g., VMware vNetwork distributed
80 switch, Cisco Nexus 1000V) provide a centralized way to configure and
81 monitor the network state of VMs that are spread across many physical
82 hosts. Open vSwitch is not a distributed vswitch itself, rather it
83 runs on each physical host and supports remote management in a way
84 that makes it easier for developers of virtualization/cloud
85 management platforms to offer distributed vswitch capabilities.
87 To aid in distribution, Open vSwitch provides two open protocols that
88 are specially designed for remote management in virtualized network
89 environments: OpenFlow, which exposes flow-based forwarding state,
90 and the OVSDB management protocol, which exposes switch port state.
91 In addition to the switch implementation itself, Open vSwitch
92 includes tools (ovs-ofctl, ovs-vsctl) that developers can script and
93 extend to provide distributed vswitch capabilities that are closely
94 integrated with their virtualization management platform.
96 Q: Why doesn't Open vSwitch support distribution?
98 A: Open vSwitch is intended to be a useful component for building
99 flexible network infrastructure. There are many different approaches
100 to distribution which balance trade-offs between simplicity,
101 scalability, hardware compatibility, convergence times, logical
102 forwarding model, etc. The goal of Open vSwitch is to be able to
103 support all as a primitive building block rather than choose a
104 particular point in the distributed design space.
106 Q: How can I contribute to the Open vSwitch Community?
108 A: You can start by joining the mailing lists and helping to answer
109 questions. You can also suggest improvements to documentation. If
110 you have a feature or bug you would like to work on, send a mail to
111 one of the mailing lists:
113 http://openvswitch.org/mlists/
119 Q: What does it mean for an Open vSwitch release to be LTS (long-term
122 A: All official releases have been through a comprehensive testing
123 process and are suitable for production use. Planned releases will
124 occur several times a year. If a significant bug is identified in an
125 LTS release, we will provide an updated release that includes the
126 fix. Releases that are not LTS may not be fixed and may just be
127 supplanted by the next major release. The current LTS release is
130 Q: What Linux kernel versions does each Open vSwitch release work with?
132 A: The following table lists the Linux kernel versions against which the
133 given versions of the Open vSwitch kernel module will successfully
134 build. The Linux kernel versions are upstream kernel versions, so
135 Linux kernels modified from the upstream sources may not build in
136 some cases even if they are based on a supported version. This is
137 most notably true of Red Hat Enterprise Linux (RHEL) kernels, which
138 are extensively modified from upstream.
140 Open vSwitch Linux kernel
141 ------------ -------------
154 Open vSwitch userspace should also work with the Linux kernel module
155 built into Linux 3.3 and later.
157 Open vSwitch userspace is not sensitive to the Linux kernel version.
158 It should build against almost any kernel, certainly against 2.6.32
161 Q: What Linux kernel versions does IPFIX flow monitoring work with?
163 A: IPFIX flow monitoring requires the Linux kernel module from Open
164 vSwitch version 1.10.90 or later.
166 Q: Should userspace or kernel be upgraded first to minimize downtime?
168 In general, the Open vSwitch userspace should be used with the
169 kernel version included in the same release or with the version
170 from upstream Linux. However, when upgrading between two releases
171 of Open vSwitch it is best to migrate userspace first to reduce
172 the possibility of incompatibilities.
174 Q: What features are not available in the Open vSwitch kernel datapath
175 that ships as part of the upstream Linux kernel?
177 A: The kernel module in upstream Linux does not include support for
178 LISP. Work is in progress to add support for LISP to the upstream
179 Linux version of the Open vSwitch kernel module. For now, if you
180 need this feature, use the kernel module from the Open vSwitch
181 distribution instead of the upstream Linux kernel module.
183 Certain features require kernel support to function or to have
184 reasonable performance. If the ovs-vswitchd log file indicates that
185 a feature is not supported, consider upgrading to a newer upstream
186 Linux release or using the kernel module paired with the userspace
189 Q: What features are not available when using the userspace datapath?
191 A: Tunnel virtual ports are not supported, as described in the
192 previous answer. It is also not possible to use queue-related
193 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
194 may not be transmitted.
196 Q: What happened to the bridge compatibility feature?
198 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
199 When it was enabled, Open vSwitch imitated the interface of the
200 Linux kernel "bridge" module. This allowed users to drop Open
201 vSwitch into environments designed to use the Linux kernel bridge
202 module without adapting the environment to use Open vSwitch.
204 Open vSwitch 1.10 and later do not support bridge compatibility.
205 The feature was dropped because version 1.10 adopted a new internal
206 architecture that made bridge compatibility difficult to maintain.
207 Now that many environments use OVS directly, it would be rarely
210 To use bridge compatibility, install OVS 1.9 or earlier, including
211 the accompanying kernel modules (both the main and bridge
212 compatibility modules), following the instructions that come with
213 the release. Be sure to start the ovs-brcompatd daemon.
219 Q: I thought Open vSwitch was a virtual Ethernet switch, but the
220 documentation keeps talking about bridges. What's a bridge?
222 A: In networking, the terms "bridge" and "switch" are synonyms. Open
223 vSwitch implements an Ethernet switch, which means that it is also
228 A: See the "VLAN" section below.
234 Q: How do I configure a port as an access port?
236 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
237 the following commands configure br0 with eth0 as a trunk port (the
238 default) and tap0 as an access port for VLAN 9:
241 ovs-vsctl add-port br0 eth0
242 ovs-vsctl add-port br0 tap0 tag=9
244 If you want to configure an already added port as an access port,
245 use "ovs-vsctl set", e.g.:
247 ovs-vsctl set port tap0 tag=9
249 Q: How do I configure a port as a SPAN port, that is, enable mirroring
250 of all traffic to that port?
252 A: The following commands configure br0 with eth0 and tap0 as trunk
253 ports. All traffic coming in or going out on eth0 or tap0 is also
254 mirrored to tap1; any traffic arriving on tap1 is dropped:
257 ovs-vsctl add-port br0 eth0
258 ovs-vsctl add-port br0 tap0
259 ovs-vsctl add-port br0 tap1 \
260 -- --id=@p get port tap1 \
261 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
262 -- set bridge br0 mirrors=@m
264 To later disable mirroring, run:
266 ovs-vsctl clear bridge br0 mirrors
268 Q: Does Open vSwitch support configuring a port in promiscuous mode?
270 A: Yes. How you configure it depends on what you mean by "promiscuous
273 - Conventionally, "promiscuous mode" is a feature of a network
274 interface card. Ordinarily, a NIC passes to the CPU only the
275 packets actually destined to its host machine. It discards
276 the rest to avoid wasting memory and CPU cycles. When
277 promiscuous mode is enabled, however, it passes every packet
278 to the CPU. On an old-style shared-media or hub-based
279 network, this allows the host to spy on all packets on the
280 network. But in the switched networks that are almost
281 everywhere these days, promiscuous mode doesn't have much
282 effect, because few packets not destined to a host are
283 delivered to the host's NIC.
285 This form of promiscuous mode is configured in the guest OS of
286 the VMs on your bridge, e.g. with "ifconfig".
288 - The VMware vSwitch uses a different definition of "promiscuous
289 mode". When you configure promiscuous mode on a VMware vNIC,
290 the vSwitch sends a copy of every packet received by the
291 vSwitch to that vNIC. That has a much bigger effect than just
292 enabling promiscuous mode in a guest OS. Rather than getting
293 a few stray packets for which the switch does not yet know the
294 correct destination, the vNIC gets every packet. The effect
295 is similar to replacing the vSwitch by a virtual hub.
297 This "promiscuous mode" is what switches normally call "port
298 mirroring" or "SPAN". For information on how to configure
299 SPAN, see "How do I configure a port as a SPAN port, that is,
300 enable mirroring of all traffic to that port?"
302 Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable
303 mirroring of all traffic to that VLAN?
305 A: The following commands configure br0 with eth0 as a trunk port and
306 tap0 as an access port for VLAN 10. All traffic coming in or going
307 out on tap0, as well as traffic coming in or going out on eth0 in
308 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
309 VLAN 10, in cases where one is present, is dropped as part of
313 ovs-vsctl add-port br0 eth0
314 ovs-vsctl add-port br0 tap0 tag=10
316 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
318 -- set bridge br0 mirrors=@m
320 To later disable mirroring, run:
322 ovs-vsctl clear bridge br0 mirrors
324 Mirroring to a VLAN can disrupt a network that contains unmanaged
325 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
326 GRE tunnel has fewer caveats than mirroring to a VLAN and should
327 generally be preferred.
329 Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
331 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
332 of the specified output-vlan. This loss of information may make
333 the mirrored traffic too hard to interpret.
335 To mirror multiple VLANs, use the commands above, but specify a
336 comma-separated list of VLANs as the value for select-vlan. To
337 mirror every VLAN, use the commands above, but omit select-vlan and
340 When a packet arrives on a VLAN that is used as a mirror output
341 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
342 floods the packet across all the ports for which the mirror output
343 VLAN is configured. (If an OpenFlow controller is in use, then it
344 can override this behavior through the flow table.) If OVS is used
345 as an intermediate switch, rather than an edge switch, this ensures
346 that the RSPAN traffic is distributed through the network.
348 Mirroring to a VLAN can disrupt a network that contains unmanaged
349 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
350 GRE tunnel has fewer caveats than mirroring to a VLAN and should
351 generally be preferred.
353 Q: How do I configure mirroring of all traffic to a GRE tunnel?
355 A: The following commands configure br0 with eth0 and tap0 as trunk
356 ports. All traffic coming in or going out on eth0 or tap0 is also
357 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
358 traffic arriving on gre0 is dropped:
361 ovs-vsctl add-port br0 eth0
362 ovs-vsctl add-port br0 tap0
363 ovs-vsctl add-port br0 gre0 \
364 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
365 -- --id=@p get port gre0 \
366 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
367 -- set bridge br0 mirrors=@m
369 To later disable mirroring and destroy the GRE tunnel:
371 ovs-vsctl clear bridge br0 mirrors
372 ovs-vcstl del-port br0 gre0
374 Q: Does Open vSwitch support ERSPAN?
376 A: No. ERSPAN is an undocumented proprietary protocol. As an
377 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
380 Q: How do I connect two bridges?
382 A: First, why do you want to do this? Two connected bridges are not
383 much different from a single bridge, so you might as well just have
384 a single bridge with all your ports on it.
386 If you still want to connect two bridges, you can use a pair of
387 patch ports. The following example creates bridges br0 and br1,
388 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
389 and br1 with a pair of patch ports.
392 ovs-vsctl add-port br0 eth0
393 ovs-vsctl add-port br0 tap0
395 ovs-vsctl add-port br1 tap1
397 -- add-port br0 patch0 \
398 -- set interface patch0 type=patch options:peer=patch1 \
399 -- add-port br1 patch1 \
400 -- set interface patch1 type=patch options:peer=patch0
402 Bridges connected with patch ports are much like a single bridge.
403 For instance, if the example above also added eth1 to br1, and both
404 eth0 and eth1 happened to be connected to the same next-hop switch,
405 then you could loop your network just as you would if you added
406 eth0 and eth1 to the same bridge (see the "Configuration Problems"
407 section below for more information).
409 If you are using Open vSwitch 1.9 or an earlier version, then you
410 need to be using the kernel module bundled with Open vSwitch rather
411 than the one that is integrated into Linux 3.3 and later, because
412 Open vSwitch 1.9 and earlier versions need kernel support for patch
413 ports. This also means that in Open vSwitch 1.9 and earlier, patch
414 ports will not work with the userspace datapath, only with the
418 Implementation Details
419 ----------------------
421 Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
423 A: Open vSwitch uses different kinds of flows for different purposes:
425 - OpenFlow flows are the most important kind of flow. OpenFlow
426 controllers use these flows to define a switch's policy.
427 OpenFlow flows support wildcards, priorities, and multiple
430 When in-band control is in use, Open vSwitch sets up a few
431 "hidden" flows, with priority higher than a controller or the
432 user can configure, that are not visible via OpenFlow. (See
433 the "Controller" section of the FAQ for more information
436 - The Open vSwitch software switch implementation uses a second
437 kind of flow internally. These flows, called "datapath" or
438 "kernel" flows, do not support priorities and comprise only a
439 single table, which makes them suitable for caching. (Like
440 OpenFlow flows, datapath flows do support wildcarding, in Open
441 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
442 also support different actions and number ports differently.
444 Datapath flows are an implementation detail that is subject to
445 change in future versions of Open vSwitch. Even with the
446 current version of Open vSwitch, hardware switch
447 implementations do not necessarily use this architecture.
449 Users and controllers directly control only the OpenFlow flow
450 table. Open vSwitch manages the datapath flow table itself, so
451 users should not normally be concerned with it.
453 Q: Why are there so many different ways to dump flows?
455 A: Open vSwitch has two kinds of flows (see the previous question), so
456 it has commands with different purposes for dumping each kind of
459 - "ovs-ofctl dump-flows <br>" dumps OpenFlow flows, excluding
460 hidden flows. This is the most commonly useful form of flow
461 dump. (Unlike the other commands, this should work with any
462 OpenFlow switch, not just Open vSwitch.)
464 - "ovs-appctl bridge/dump-flows <br>" dumps OpenFlow flows,
465 including hidden flows. This is occasionally useful for
466 troubleshooting suspected issues with in-band control.
468 - "ovs-dpctl dump-flows [dp]" dumps the datapath flow table
469 entries for a Linux kernel-based datapath. In Open vSwitch
470 1.10 and later, ovs-vswitchd merges multiple switches into a
471 single datapath, so it will show all the flows on all your
472 kernel-based switches. This command can occasionally be
473 useful for debugging.
475 - "ovs-appctl dpif/dump-flows <br>", new in Open vSwitch 1.10,
476 dumps datapath flows for only the specified bridge, regardless
483 Q: I just upgraded and I see a performance drop. Why?
485 A: The OVS kernel datapath may have been updated to a newer version than
486 the OVS userspace components. Sometimes new versions of OVS kernel
487 module add functionality that is backwards compatible with older
488 userspace components but may cause a drop in performance with them.
489 Especially, if a kernel module from OVS 2.1 or newer is paired with
490 OVS userspace 1.10 or older, there will be a performance drop for
493 Updating the OVS userspace components to the latest released
494 version should fix the performance degradation.
496 To get the best possible performance and functionality, it is
497 recommended to pair the same versions of the kernel module and OVS
501 Configuration Problems
502 ----------------------
504 Q: I created a bridge and added my Ethernet port to it, using commands
508 ovs-vsctl add-port br0 eth0
510 and as soon as I ran the "add-port" command I lost all connectivity
513 A: A physical Ethernet device that is part of an Open vSwitch bridge
514 should not have an IP address. If one does, then that IP address
515 will not be fully functional.
517 You can restore functionality by moving the IP address to an Open
518 vSwitch "internal" device, such as the network device named after
519 the bridge itself. For example, assuming that eth0's IP address is
520 192.168.128.5, you could run the commands below to fix up the
523 ifconfig eth0 0.0.0.0
524 ifconfig br0 192.168.128.5
526 (If your only connection to the machine running OVS is through the
527 IP address in question, then you would want to run all of these
528 commands on a single command line, or put them into a script.) If
529 there were any additional routes assigned to eth0, then you would
530 also want to use commands to adjust these routes to go through br0.
532 If you use DHCP to obtain an IP address, then you should kill the
533 DHCP client that was listening on the physical Ethernet interface
534 (e.g. eth0) and start one listening on the internal interface
535 (e.g. br0). You might still need to manually clear the IP address
536 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
538 There is no compelling reason why Open vSwitch must work this way.
539 However, this is the way that the Linux kernel bridge module has
540 always worked, so it's a model that those accustomed to Linux
541 bridging are already used to. Also, the model that most people
542 expect is not implementable without kernel changes on all the
543 versions of Linux that Open vSwitch supports.
545 By the way, this issue is not specific to physical Ethernet
546 devices. It applies to all network devices except Open vswitch
549 Q: I created a bridge and added a couple of Ethernet ports to it,
550 using commands like these:
553 ovs-vsctl add-port br0 eth0
554 ovs-vsctl add-port br0 eth1
556 and now my network seems to have melted: connectivity is unreliable
557 (even connectivity that doesn't go through Open vSwitch), all the
558 LEDs on my physical switches are blinking, wireshark shows
559 duplicated packets, and CPU usage is very high.
561 A: More than likely, you've looped your network. Probably, eth0 and
562 eth1 are connected to the same physical Ethernet switch. This
563 yields a scenario where OVS receives a broadcast packet on eth0 and
564 sends it out on eth1, then the physical switch connected to eth1
565 sends the packet back on eth0, and so on forever. More complicated
566 scenarios, involving a loop through multiple switches, are possible
569 The solution depends on what you are trying to do:
571 - If you added eth0 and eth1 to get higher bandwidth or higher
572 reliability between OVS and your physical Ethernet switch,
573 use a bond. The following commands create br0 and then add
574 eth0 and eth1 as a bond:
577 ovs-vsctl add-bond br0 bond0 eth0 eth1
579 Bonds have tons of configuration options. Please read the
580 documentation on the Port table in ovs-vswitchd.conf.db(5)
583 - Perhaps you don't actually need eth0 and eth1 to be on the
584 same bridge. For example, if you simply want to be able to
585 connect each of them to virtual machines, then you can put
586 each of them on a bridge of its own:
589 ovs-vsctl add-port br0 eth0
592 ovs-vsctl add-port br1 eth1
594 and then connect VMs to br0 and br1. (A potential
595 disadvantage is that traffic cannot directly pass between br0
596 and br1. Instead, it will go out eth0 and come back in eth1,
599 - If you have a redundant or complex network topology and you
600 want to prevent loops, turn on spanning tree protocol (STP).
601 The following commands create br0, enable STP, and add eth0
602 and eth1 to the bridge. The order is important because you
603 don't want have to have a loop in your network even
607 ovs-vsctl set bridge br0 stp_enable=true
608 ovs-vsctl add-port br0 eth0
609 ovs-vsctl add-port br0 eth1
611 The Open vSwitch implementation of STP is not well tested.
612 Please report any bugs you observe, but if you'd rather avoid
613 acting as a beta tester then another option might be your
616 Q: I can't seem to use Open vSwitch in a wireless network.
618 A: Wireless base stations generally only allow packets with the source
619 MAC address of NIC that completed the initial handshake.
620 Therefore, without MAC rewriting, only a single device can
621 communicate over a single wireless link.
623 This isn't specific to Open vSwitch, it's enforced by the access
624 point, so the same problems will show up with the Linux bridge or
625 any other way to do bridging.
627 Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
629 A: PPP most commonly carries IP packets, but Open vSwitch works only
630 with Ethernet frames. The correct way to interface PPP to an
631 Ethernet network is usually to use routing instead of switching.
633 Q: Is there any documentation on the database tables and fields?
635 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
637 Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
638 I only see a datapath called "ovs-system". How can I see datapath
639 information about a particular bridge?
641 A: In version 1.9.0, OVS switched to using a single datapath that is
642 shared by all bridges of that type. The "ovs-appctl dpif/*"
643 commands provide similar functionality that is scoped by the bridge.
645 Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
646 see the port in the datapath?
648 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
649 cannot be loaded at the same time. It is likely that on your system the
650 Linux GRE module is already loaded and blocking OVS (to confirm, check
651 dmesg for errors regarding GRE registration). To fix this, unload all
652 GRE modules that appear in lsmod as well as the OVS kernel module. You
653 can then reload the OVS module following the directions in INSTALL,
654 which will ensure that dependencies are satisfied.
657 Quality of Service (QoS)
658 ------------------------
660 Q: How do I configure Quality of Service (QoS)?
662 A: Suppose that you want to set up bridge br0 connected to physical
663 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
664 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
665 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
666 could configure the bridge this way:
670 add-port br0 eth0 -- \
671 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
672 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
673 set port eth0 qos=@newqos -- \
674 --id=@newqos create qos type=linux-htb \
675 other-config:max-rate=1000000000 \
676 queues:123=@vif10queue \
677 queues:234=@vif20queue -- \
678 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
679 --id=@vif20queue create queue other-config:max-rate=20000000
681 At this point, bridge br0 is configured with the ports and eth0 is
682 configured with the queues that you need for QoS, but nothing is
683 actually directing packets from vif1.0 or vif2.0 to the queues that
684 we have set up for them. That means that all of the packets to
685 eth0 are going to the "default queue", which is not what we want.
687 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
688 queues reserved for them:
690 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
691 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
693 Each of the above flows matches on the input port, sets up the
694 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
695 executes the "normal" action, which performs the same switching
696 that Open vSwitch would have done without any OpenFlow flows being
697 present. (We know that vif1.0 and vif2.0 have OpenFlow port
698 numbers 5 and 6, respectively, because we set their ofport_request
699 columns above. If we had not done that, then we would have needed
700 to find out their port numbers before setting up these flows.)
702 Now traffic going from vif1.0 or vif2.0 to eth0 should be
705 By the way, if you delete the bridge created by the above commands,
710 then that will leave one unreferenced QoS record and two
711 unreferenced Queue records in the Open vSwich database. One way to
712 clear them out, assuming you don't have other QoS or Queue records
713 that you want to keep, is:
715 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
717 If you do want to keep some QoS or Queue records, or the Open
718 vSwitch you are using is older than version 1.8 (which added the
719 --all option), then you will have to destroy QoS and Queue records
722 Q: I configured Quality of Service (QoS) in my OpenFlow network by
723 adding records to the QoS and Queue table, but the results aren't
726 A: Did you install OpenFlow flows that use your queues? This is the
727 primary way to tell Open vSwitch which queues you want to use. If
728 you don't do this, then the default queue will be used, which will
729 probably not have the effect you want.
731 Refer to the previous question for an example.
733 Q: I'd like to take advantage of some QoS feature that Open vSwitch
734 doesn't yet support. How do I do that?
736 A: Open vSwitch does not implement QoS itself. Instead, it can
737 configure some, but not all, of the QoS features built into the
738 Linux kernel. If you need some QoS feature that OVS cannot
739 configure itself, then the first step is to figure out whether
740 Linux QoS supports that feature. If it does, then you can submit a
741 patch to support Open vSwitch configuration for that feature, or
742 you can use "tc" directly to configure the feature in Linux. (If
743 Linux QoS doesn't support the feature you want, then first you have
744 to add that support to Linux.)
746 Q: I configured QoS, correctly, but my measurements show that it isn't
747 working as well as I expect.
749 A: With the Linux kernel, the Open vSwitch implementation of QoS has
752 - Open vSwitch configures a subset of Linux kernel QoS
753 features, according to what is in OVSDB. It is possible that
754 this code has bugs. If you believe that this is so, then you
755 can configure the Linux traffic control (QoS) stack directly
756 with the "tc" program. If you get better results that way,
757 you can send a detailed bug report to bugs@openvswitch.org.
759 It is certain that Open vSwitch cannot configure every Linux
760 kernel QoS feature. If you need some feature that OVS cannot
761 configure, then you can also use "tc" directly (or add that
764 - The Open vSwitch implementation of OpenFlow allows flows to
765 be directed to particular queues. This is pretty simple and
766 unlikely to have serious bugs at this point.
768 However, most problems with QoS on Linux are not bugs in Open
769 vSwitch at all. They tend to be either configuration errors
770 (please see the earlier questions in this section) or issues with
771 the traffic control (QoS) stack in Linux. The Open vSwitch
772 developers are not experts on Linux traffic control. We suggest
773 that, if you believe you are encountering a problem with Linux
774 traffic control, that you consult the tc manpages (e.g. tc(8),
775 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
776 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
778 Q: Does Open vSwitch support OpenFlow meters?
780 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
781 OpenFlow meters. There is no implementation of meters in the Open
782 vSwitch software switch (neither the kernel-based nor userspace
791 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
792 partition a single switch into multiple switches. Suppose, for
793 example, that you have two groups of machines, group A and group B.
794 You want the machines in group A to be able to talk to each other,
795 and you want the machine in group B to be able to talk to each
796 other, but you don't want the machines in group A to be able to
797 talk to the machines in group B. You can do this with two
798 switches, by plugging the machines in group A into one switch and
799 the machines in group B into the other switch.
801 If you only have one switch, then you can use VLANs to do the same
802 thing, by configuring the ports for machines in group A as VLAN
803 "access ports" for one VLAN and the ports for group B as "access
804 ports" for a different VLAN. The switch will only forward packets
805 between ports that are assigned to the same VLAN, so this
806 effectively subdivides your single switch into two independent
807 switches, one for each group of machines.
809 So far we haven't said anything about VLAN headers. With access
810 ports, like we've described so far, no VLAN header is present in
811 the Ethernet frame. This means that the machines (or switches)
812 connected to access ports need not be aware that VLANs are
813 involved, just like in the case where we use two different physical
816 Now suppose that you have a whole bunch of switches in your
817 network, instead of just one, and that some machines in group A are
818 connected directly to both switches 1 and 2. To allow these
819 machines to talk to each other, you could add an access port for
820 group A's VLAN to switch 1 and another to switch 2, and then
821 connect an Ethernet cable between those ports. That works fine,
822 but it doesn't scale well as the number of switches and the number
823 of VLANs increases, because you use up a lot of valuable switch
824 ports just connecting together your VLANs.
826 This is where VLAN headers come in. Instead of using one cable and
827 two ports per VLAN to connect a pair of switches, we configure a
828 port on each switch as a VLAN "trunk port". Packets sent and
829 received on a trunk port carry a VLAN header that says what VLAN
830 the packet belongs to, so that only two ports total are required to
831 connect the switches, regardless of the number of VLANs in use.
832 Normally, only switches (either physical or virtual) are connected
833 to a trunk port, not individual hosts, because individual hosts
834 don't expect to see a VLAN header in the traffic that they receive.
836 None of the above discussion says anything about particular VLAN
837 numbers. This is because VLAN numbers are completely arbitrary.
838 One must only ensure that a given VLAN is numbered consistently
839 throughout a network and that different VLANs are given different
840 numbers. (That said, VLAN 0 is usually synonymous with a packet
841 that has no VLAN header, and VLAN 4095 is reserved.)
845 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
846 bugs. If you are having problems with VLANs that you suspect to be
847 driver related, then you have several options:
849 - Upgrade to Linux 3.3 or later.
851 - Build and install a fixed version of the particular driver
852 that is causing trouble, if one is available.
854 - Use a NIC whose driver does not have VLAN problems.
856 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
857 that works around bugs in kernel drivers. To enable VLAN
858 splinters on interface eth0, use the command:
860 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
862 For VLAN splinters to be effective, Open vSwitch must know
863 which VLANs are in use. See the "VLAN splinters" section in
864 the Interface table in ovs-vswitchd.conf.db(5) for details on
865 how Open vSwitch infers in-use VLANs.
867 VLAN splinters increase memory use and reduce performance, so
868 use them only if needed.
870 - Apply the "vlan workaround" patch from the XenServer kernel
871 patch queue, build Open vSwitch against this patched kernel,
872 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
873 workaround for each interface whose driver is buggy.
875 (This is a nontrivial exercise, so this option is included
876 only for completeness.)
878 It is not always easy to tell whether a Linux kernel driver has
879 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
880 can help you test. See their manpages for details. Of the two
881 utilities, ovs-test(8) is newer and more thorough, but
882 ovs-vlan-test(8) may be easier to use.
884 Q: VLANs still don't work. I've tested the driver so I know that it's OK.
886 A: Do you have VLANs enabled on the physical switch that OVS is
887 attached to? Make sure that the port is configured to trunk the
888 VLAN or VLANs that you are using with OVS.
890 Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
891 and to its destination host, but OVS seems to drop incoming return
894 A: It's possible that you have the VLAN configured on your physical
895 switch as the "native" VLAN. In this mode, the switch treats
896 incoming packets either tagged with the native VLAN or untagged as
897 part of the native VLAN. It may also send outgoing packets in the
898 native VLAN without a VLAN tag.
900 If this is the case, you have two choices:
902 - Change the physical switch port configuration to tag packets
903 it forwards to OVS with the native VLAN instead of forwarding
906 - Change the OVS configuration for the physical port to a
907 native VLAN mode. For example, the following sets up a
908 bridge with port eth0 in "native-tagged" mode in VLAN 9:
911 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
913 In this situation, "native-untagged" mode will probably work
914 equally well. Refer to the documentation for the Port table
915 in ovs-vswitchd.conf.db(5) for more information.
917 Q: I added a pair of VMs on different VLANs, like this:
920 ovs-vsctl add-port br0 eth0
921 ovs-vsctl add-port br0 tap0 tag=9
922 ovs-vsctl add-port br0 tap1 tag=10
924 but the VMs can't access each other, the external network, or the
927 A: It is to be expected that the VMs can't access each other. VLANs
928 are a means to partition a network. When you configured tap0 and
929 tap1 as access ports for different VLANs, you indicated that they
930 should be isolated from each other.
932 As for the external network and the Internet, it seems likely that
933 the machines you are trying to access are not on VLAN 9 (or 10) and
934 that the Internet is not available on VLAN 9 (or 10).
936 Q: I added a pair of VMs on the same VLAN, like this:
939 ovs-vsctl add-port br0 eth0
940 ovs-vsctl add-port br0 tap0 tag=9
941 ovs-vsctl add-port br0 tap1 tag=9
943 The VMs can access each other, but not the external network or the
946 A: It seems likely that the machines you are trying to access in the
947 external network are not on VLAN 9 and that the Internet is not
948 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
949 trunk VLAN on the upstream switch port to which eth0 is connected.
951 Q: Can I configure an IP address on a VLAN?
953 A: Yes. Use an "internal port" configured as an access port. For
954 example, the following configures IP address 192.168.0.7 on VLAN 9.
955 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
956 they have an 802.1Q header with VLAN 9. Conversely, traffic
957 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
961 ovs-vsctl add-port br0 eth0
962 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
963 ifconfig vlan9 192.168.0.7
965 See also the following question.
967 Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
971 ovs-vsctl add-port br0 eth0
972 ifconfig br0 192.168.0.5
973 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
974 ifconfig vlan9 192.168.0.9
976 but other hosts that are only on VLAN 0 can reach the IP address
977 configured on VLAN 9. What's going on?
979 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
980 approaches to IP address handling in Internet hosts:
982 - In the "Strong ES Model", where an ES is a host ("End
983 System"), an IP address is primarily associated with a
984 particular interface. The host discards packets that arrive
985 on interface A if they are destined for an IP address that is
986 configured on interface B. The host never sends packets from
987 interface A using a source address configured on interface B.
989 - In the "Weak ES Model", an IP address is primarily associated
990 with a host. The host accepts packets that arrive on any
991 interface if they are destined for any of the host's IP
992 addresses, even if the address is configured on some
993 interface other than the one on which it arrived. The host
994 does not restrict itself to sending packets from an IP
995 address associated with the originating interface.
997 Linux uses the weak ES model. That means that when packets
998 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
999 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1000 address, even though they were not received on vlan9, the network
1003 To simulate the strong ES model on Linux, one may add iptables rule
1004 to filter packets based on source and destination address and
1005 adjust ARP configuration with sysctls.
1007 BSD uses the strong ES model.
1009 Q: My OpenFlow controller doesn't see the VLANs that I expect.
1011 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1012 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1013 implementation of the OpenFlow "normal switching" action. By
1014 default, when Open vSwitch isn't connected to a controller and
1015 nothing has been manually configured in the flow table, all traffic
1016 goes through the "normal switching" action. But, if you set up
1017 OpenFlow flows on your own, through a controller or using ovs-ofctl
1018 or through other means, then you have to implement VLAN handling
1021 You can use "normal switching" as a component of your OpenFlow
1022 actions, e.g. by putting "normal" into the lists of actions on
1023 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1024 controller. In situations where this is not suitable, you can
1025 implement VLAN handling yourself, e.g.:
1027 - If a packet comes in on an access port, and the flow table
1028 needs to send it out on a trunk port, then the flow can add
1029 the appropriate VLAN tag with the "mod_vlan_vid" action.
1031 - If a packet comes in on a trunk port, and the flow table
1032 needs to send it out on an access port, then the flow can
1033 strip the VLAN tag with the "strip_vlan" action.
1035 Q: I configured ports on a bridge as access ports with different VLAN
1038 ovs-vsctl add-br br0
1039 ovs-vsctl set-controller br0 tcp:192.168.0.10:6633
1040 ovs-vsctl add-port br0 eth0
1041 ovs-vsctl add-port br0 tap0 tag=9
1042 ovs-vsctl add-port br0 tap1 tag=10
1044 but the VMs running behind tap0 and tap1 can still communicate,
1045 that is, they are not isolated from each other even though they are
1048 A: Do you have a controller configured on br0 (as the commands above
1049 do)? If so, then this is a variant on the previous question, "My
1050 OpenFlow controller doesn't see the VLANs that I expect," and you
1051 can refer to the answer there for more information.
1059 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1060 to solve the scaling challenges of VLAN networks in a multi-tenant
1061 environment. VXLAN is an overlay network which transports an L2 network
1062 over an existing L3 network. For more information on VXLAN, please see
1063 the IETF draft available here:
1065 http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-03
1067 Q: How much of the VXLAN protocol does Open vSwitch currently support?
1069 A: Open vSwitch currently supports the framing format for packets on the
1070 wire. There is currently no support for the multicast aspects of VXLAN.
1071 To get around the lack of multicast support, it is possible to
1072 pre-provision MAC to IP address mappings either manually or from a
1075 Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1078 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1079 is 4789. However, it is possible to configure the destination UDP port
1080 manually on a per-VXLAN tunnel basis. An example of this configuration is
1083 ovs-vsctl add-br br0
1084 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1085 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1086 options:dst_port=8472
1089 Using OpenFlow (Manually or Via Controller)
1090 -------------------------------------------
1092 Q: What versions of OpenFlow does Open vSwitch support?
1094 A: The following table lists the versions of OpenFlow supported by
1095 each version of Open vSwitch:
1097 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4
1098 =============== ===== ===== ===== ===== =====
1099 1.9 and earlier yes --- --- --- ---
1100 1.10 yes --- [*] [*] ---
1101 1.11 yes --- [*] [*] ---
1102 2.0 yes [*] [*] [*] ---
1103 2.1 yes [*] [*] [*] ---
1104 2.2 yes [*] [*] [*] [%]
1106 [*] Supported, with one or more missing features.
1107 [%] Support is unsafe: ovs-vswitchd will abort when certain
1108 unimplemented features are tested.
1110 Because of missing features, OpenFlow 1.1, 1.2, and 1.3 must be
1111 enabled manually. The following command enables OpenFlow 1.0, 1.1,
1112 1.2, and 1.3 on bridge br0:
1114 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1116 Use the -O option to enable support for later versions of OpenFlow
1117 in ovs-ofctl. For example:
1119 ovs-ofctl -O OpenFlow13 dump-flows br0
1121 OpenFlow 1.4 is a special case, because it is not implemented
1122 safely: ovs-vswitchd will abort when certain unimplemented features
1123 are tested. Thus, for now it is suitable only for experimental
1124 use. ovs-vswitchd will only allow OpenFlow 1.4 to be enabled
1125 (which must be done in the same way described above) when it is
1126 invoked with a special --enable-of14 command line option.
1128 OPENFLOW-1.1+ in the Open vSwitch source tree tracks support for
1129 OpenFlow 1.1 and later features. When support for a given OpenFlow
1130 version is solidly implemented, Open vSwitch will enable that
1133 Q: Does Open vSwitch support MPLS?
1135 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1136 these versions can match on MPLS Ethernet types, but they cannot
1137 match, push, or pop MPLS labels, nor can they look past MPLS labels
1138 into the encapsulated packet.
1140 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1141 for MPLS. With the userspace datapath only, these versions can
1142 match, push, or pop a single MPLS label, but they still cannot look
1143 past MPLS labels (even after popping them) into the encapsulated
1144 packet. Kernel datapath support is unchanged from earlier
1147 Open vSwitch version 2.2 will be able to match, push, or pop up to
1148 3 MPLS labels. Looking past MPLS labels into the encapsulated
1149 packet will still be unsupported. Both userspace and kernel
1150 datapaths will be supported, but MPLS processing always happens in
1151 userspace either way, so kernel datapath performance will be
1154 Q: I'm getting "error type 45250 code 0". What's that?
1156 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1157 vSwitch uses this extension when it must report an error to an
1158 OpenFlow controller but no standard OpenFlow error code is
1161 Open vSwitch logs the errors that it sends to controllers, so the
1162 easiest thing to do is probably to look at the ovs-vswitchd log to
1163 find out what the error was.
1165 If you want to dissect the extended error message yourself, the
1166 format is documented in include/openflow/nicira-ext.h in the Open
1167 vSwitch source distribution. The extended error codes are
1168 documented in lib/ofp-errors.h.
1170 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1171 doesn't actually appear through the OpenFlow connection, even
1172 though I know that it's going through.
1173 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1174 to apply to certain traffic, especially traffic between OVS and
1175 the controller itself.
1177 A: By default, Open vSwitch assumes that OpenFlow controllers are
1178 connected "in-band", that is, that the controllers are actually
1179 part of the network that is being controlled. In in-band mode,
1180 Open vSwitch sets up special "hidden" flows to make sure that
1181 traffic can make it back and forth between OVS and the controllers.
1182 These hidden flows are higher priority than any flows that can be
1183 set up through OpenFlow, and they are not visible through normal
1184 OpenFlow flow table dumps.
1186 Usually, the hidden flows are desirable and helpful, but
1187 occasionally they can cause unexpected behavior. You can view the
1188 full OpenFlow flow table, including hidden flows, on bridge br0
1191 ovs-appctl bridge/dump-flows br0
1193 to help you debug. The hidden flows are those with priorities
1194 greater than 65535 (the maximum priority that can be set with
1197 The DESIGN file at the top level of the Open vSwitch source
1198 distribution describes the in-band model in detail.
1200 If your controllers are not actually in-band (e.g. they are on
1201 localhost via 127.0.0.1, or on a separate network), then you should
1202 configure your controllers in "out-of-band" mode. If you have one
1203 controller on bridge br0, then you can configure out-of-band mode
1206 ovs-vsctl set controller br0 connection-mode=out-of-band
1208 Q: I configured all my controllers for out-of-band control mode but
1209 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1211 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1212 set-manager"). By default, Open vSwitch assumes that managers need
1213 in-band rules set up on every bridge. You can disable these rules
1216 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1218 This actually disables in-band control entirely for the bridge, as
1219 if all the bridge's controllers were configured for out-of-band
1222 Q: My OpenFlow controller doesn't see the VLANs that I expect.
1224 A: See answer under "VLANs", above.
1226 Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1227 but I got a funny message like this:
1229 ofp_util|INFO|normalization changed ofp_match, details:
1230 ofp_util|INFO| pre: nw_dst=192.168.0.1
1233 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1234 match had disappeared, so that the flow ends up matching every
1237 A: The term "normalization" in the log message means that a flow
1238 cannot match on an L3 field without saying what L3 protocol is in
1239 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1240 so the L3 field match was dropped.
1242 In this case, the L3 protocol could be IP or ARP. A correct
1243 command for each possibility is, respectively:
1245 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1249 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1251 Similarly, a flow cannot match on an L4 field without saying what
1252 L4 protocol is in use. For example, the flow match "tp_src=1234"
1253 is, by itself, meaningless and will be ignored. Instead, to match
1254 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1255 source port 1234, write "udp,tp_src=1234".
1257 Q: How can I figure out the OpenFlow port number for a given port?
1259 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1260 respond with an OFPT_FEATURES_REPLY that, among other information,
1261 includes a mapping between OpenFlow port names and numbers. From a
1262 command prompt, "ovs-ofctl show br0" makes such a request and
1263 prints the response for switch br0.
1265 The Interface table in the Open vSwitch database also maps OpenFlow
1266 port names to numbers. To print the OpenFlow port number
1267 associated with interface eth0, run:
1269 ovs-vsctl get Interface eth0 ofport
1271 You can print the entire mapping with:
1273 ovs-vsctl -- --columns=name,ofport list Interface
1275 but the output mixes together interfaces from all bridges in the
1276 database, so it may be confusing if more than one bridge exists.
1278 In the Open vSwitch database, ofport value -1 means that the
1279 interface could not be created due to an error. (The Open vSwitch
1280 log should indicate the reason.) ofport value [] (the empty set)
1281 means that the interface hasn't been created yet. The latter is
1282 normally an intermittent condition (unless ovs-vswitchd is not
1285 Q: I added some flows with my controller or with ovs-ofctl, but when I
1286 run "ovs-dpctl dump-flows" I don't see them.
1288 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1289 won't display the information that you want. You want to use
1290 "ovs-ofctl dump-flows" instead.
1292 Q: It looks like each of the interfaces in my bonded port shows up
1293 as an individual OpenFlow port. Is that right?
1295 A: Yes, Open vSwitch makes individual bond interfaces visible as
1296 OpenFlow ports, rather than the bond as a whole. The interfaces
1297 are treated together as a bond for only a few purposes:
1299 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1300 controller is not configured, this happens implicitly to
1303 - Mirrors configured for output to a bonded port.
1305 It would make a lot of sense for Open vSwitch to present a bond as
1306 a single OpenFlow port. If you want to contribute an
1307 implementation of such a feature, please bring it up on the Open
1308 vSwitch development mailing list at dev@openvswitch.org.
1310 Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1311 multiple hosts, and other components. The behavior isn't what I
1314 A: To debug network behavior problems, trace the path of a packet,
1315 hop-by-hop, from its origin in one host to a remote host. If
1316 that's correct, then trace the path of the response packet back to
1319 Usually a simple ICMP echo request and reply ("ping") packet is
1320 good enough. Start by initiating an ongoing "ping" from the origin
1321 host to a remote host. If you are tracking down a connectivity
1322 problem, the "ping" will not display any successful output, but
1323 packets are still being sent. (In this case the packets being sent
1324 are likely ARP rather than ICMP.)
1326 Tools available for tracing include the following:
1328 - "tcpdump" and "wireshark" for observing hops across network
1329 devices, such as Open vSwitch internal devices and physical
1332 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1333 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1334 These tools allow one to observe the actions being taken on
1335 packets in ongoing flows.
1337 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1338 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1339 documentation, and "Why are there so many different ways to
1340 dump flows?" above for some background.
1342 - "ovs-appctl ofproto/trace" to observe the logic behind how
1343 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1344 documentation. You can out more details about a given flow
1345 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1346 a flow from the output into an "ovs-appctl ofproto/trace"
1349 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1350 observe what goes on at these physical hops.
1352 Starting at the origin of a given packet, observe the packet at
1353 each hop in turn. For example, in one plausible scenario, you
1356 1. "tcpdump" the "eth" interface through which an ARP egresses
1357 a VM, from inside the VM.
1359 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1360 ingresses the host machine.
1362 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1363 the host interface through which the ARP egresses the
1364 physical machine. You may need to use "ovs-dpctl show" to
1365 interpret the port numbers. If the output seems surprising,
1366 you can use "ovs-appctl ofproto/trace" to observe details of
1367 how ovs-vswitchd determined the actions in the "ovs-dpctl
1370 4. "tcpdump" the "eth" interface through which the ARP egresses
1371 the physical machine.
1373 5. "tcpdump" the "eth" interface through which the ARP
1374 ingresses the physical machine, at the remote host that
1377 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1378 remote host that receives the ARP and observe the VM "vif"
1379 or "tap" interface to which the flow is directed. Again,
1380 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1382 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1385 8. "tcpdump" the "eth" interface through which the ARP
1386 ingresses a VM, from inside the VM.
1388 It is likely that during one of these steps you will figure out the
1389 problem. If not, then follow the ARP reply back to the origin, in
1392 Q: How do I make a flow drop packets?
1394 A: To drop a packet is to receive it without forwarding it. OpenFlow
1395 explicitly specifies forwarding actions. Thus, a flow with an
1396 empty set of actions does not forward packets anywhere, causing
1397 them to be dropped. You can specify an empty set of actions with
1398 "actions=" on the ovs-ofctl command line. For example:
1400 ovs-ofctl add-flow br0 priority=65535,actions=
1402 would cause every packet entering switch br0 to be dropped.
1404 You can write "drop" explicitly if you like. The effect is the
1405 same. Thus, the following command also causes every packet
1406 entering switch br0 to be dropped:
1408 ovs-ofctl add-flow br0 priority=65535,actions=drop
1410 "drop" is not an action, either in OpenFlow or Open vSwitch.
1411 Rather, it is only a way to say that there are no actions.
1413 Q: I added a flow to send packets out the ingress port, like this:
1415 ovs-ofctl add-flow br0 in_port=2,actions=2
1417 but OVS drops the packets instead.
1419 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1420 out its ingress port. The rationale is that dropping these packets
1421 makes it harder to loop the network. Sometimes this behavior can
1422 even be convenient, e.g. it is often the desired behavior in a flow
1423 that forwards a packet to several ports ("floods" the packet).
1425 Sometimes one really needs to send a packet out its ingress port
1426 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1427 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1429 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1431 This also works in some circumstances where the flow doesn't match
1432 on the input port. For example, if you know that your switch has
1433 five ports numbered 2 through 6, then the following will send every
1434 received packet out every port, even its ingress port:
1436 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1440 ovs-ofctl add-flow br0 actions=all,in_port
1442 Sometimes, in complicated flow tables with multiple levels of
1443 "resubmit" actions, a flow needs to output to a particular port
1444 that may or may not be the ingress port. It's difficult to take
1445 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1446 provides, as an OpenFlow extension, the ability to modify the
1447 in_port field. Whatever value is currently in the in_port field is
1448 the port to which outputs will be dropped, as well as the
1449 destination for OFPP_IN_PORT. This means that the following will
1450 reliably output to port 2 or to ports 2 through 6, respectively:
1452 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1453 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1455 If the input port is important, then one may save and restore it on
1458 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1459 load:0->NXM_OF_IN_PORT[],\
1461 pop:NXM_OF_IN_PORT[]
1463 Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1464 on port 2. I set up flows to forward only traffic destined to the
1465 other host and drop other traffic, like this:
1467 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1468 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1469 priority=0,actions=drop
1471 But it doesn't work--I don't get any connectivity when I do this.
1474 A: These flows drop the ARP packets that IP hosts use to establish IP
1475 connectivity over Ethernet. To solve the problem, add flows to
1476 allow ARP to pass between the hosts:
1478 priority=5,in_port=1,arp,actions=2
1479 priority=5,in_port=2,arp,actions=1
1481 This issue can manifest other ways, too. The following flows that
1482 match on Ethernet addresses instead of IP addresses will also drop
1483 ARP packets, because ARP requests are broadcast instead of being
1484 directed to a specific host:
1486 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1487 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1488 priority=0,actions=drop
1490 The solution already described above will also work in this case.
1491 It may be better to add flows to allow all multicast and broadcast
1494 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1495 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1501 Q: How do I implement a new OpenFlow message?
1503 A: Add your new message to "enum ofpraw" and "enum ofptype" in
1504 lib/ofp-msgs.h, following the existing pattern. Then recompile and
1505 fix all of the new warnings, implementing new functionality for the
1506 new message as needed. (If you configure with --enable-Werror, as
1507 described in INSTALL, then it is impossible to miss any warnings.)
1509 If you need to add an OpenFlow vendor extension message for a
1510 vendor that doesn't yet have any extension messages, then you will
1511 also need to edit build-aux/extract-ofp-msgs.
1517 bugs@openvswitch.org
1518 http://openvswitch.org/