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: I get an error like this when I configure Open vSwitch:
163 configure: error: Linux kernel in <dir> is version <x>, but
164 version newer than <y> is not supported (please refer to the
169 A: If there is a newer version of Open vSwitch, consider building that
170 one, because it may support the kernel that you are building
171 against. (To find out, consult the table in the previous answer.)
173 Otherwise, use the Linux kernel module supplied with the kernel
174 that you are using. All versions of Open vSwitch userspace are
175 compatible with all versions of the Open vSwitch kernel module, so
176 this will also work. See also the following question.
178 Q: What features are not available in the Open vSwitch kernel datapath
179 that ships as part of the upstream Linux kernel?
181 A: The kernel module in upstream Linux does not include support for
182 LISP. Work is in progress to add support for LISP to the upstream
183 Linux version of the Open vSwitch kernel module. For now, if you
184 need this feature, use the kernel module from the Open vSwitch
185 distribution instead of the upstream Linux kernel module.
187 Certain features require kernel support to function or to have
188 reasonable performance. If the ovs-vswitchd log file indicates that
189 a feature is not supported, consider upgrading to a newer upstream
190 Linux release or using the kernel module paired with the userspace
193 Q: Why do tunnels not work when using a kernel module other than the
194 one packaged with Open vSwitch?
196 A: Support for tunnels was added to the upstream Linux kernel module
197 after the rest of Open vSwitch. As a result, some kernels may contain
198 support for Open vSwitch but not tunnels. The minimum kernel version
199 that supports each tunnel protocol is:
201 Protocol Linux Kernel
202 -------- ------------
207 If you are using a version of the kernel that is older than the one
208 listed above, it is still possible to use that tunnel protocol. However,
209 you must compile and install the kernel module included with the Open
210 vSwitch distribution rather than the one on your machine. If problems
211 persist after doing this, check to make sure that the module that is
212 loaded is the one you expect.
214 Q: What features are not available when using the userspace datapath?
216 A: Tunnel virtual ports are not supported, as described in the
217 previous answer. It is also not possible to use queue-related
218 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
219 may not be transmitted.
221 Q: What Linux kernel versions does IPFIX flow monitoring work with?
223 A: IPFIX flow monitoring requires the Linux kernel module from Open
224 vSwitch version 1.10.90 or later.
226 Q: Should userspace or kernel be upgraded first to minimize downtime?
228 In general, the Open vSwitch userspace should be used with the
229 kernel version included in the same release or with the version
230 from upstream Linux. However, when upgrading between two releases
231 of Open vSwitch it is best to migrate userspace first to reduce
232 the possibility of incompatibilities.
234 Q: What happened to the bridge compatibility feature?
236 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
237 When it was enabled, Open vSwitch imitated the interface of the
238 Linux kernel "bridge" module. This allowed users to drop Open
239 vSwitch into environments designed to use the Linux kernel bridge
240 module without adapting the environment to use Open vSwitch.
242 Open vSwitch 1.10 and later do not support bridge compatibility.
243 The feature was dropped because version 1.10 adopted a new internal
244 architecture that made bridge compatibility difficult to maintain.
245 Now that many environments use OVS directly, it would be rarely
248 To use bridge compatibility, install OVS 1.9 or earlier, including
249 the accompanying kernel modules (both the main and bridge
250 compatibility modules), following the instructions that come with
251 the release. Be sure to start the ovs-brcompatd daemon.
257 Q: I thought Open vSwitch was a virtual Ethernet switch, but the
258 documentation keeps talking about bridges. What's a bridge?
260 A: In networking, the terms "bridge" and "switch" are synonyms. Open
261 vSwitch implements an Ethernet switch, which means that it is also
266 A: See the "VLAN" section below.
272 Q: How do I configure a port as an access port?
274 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
275 the following commands configure br0 with eth0 as a trunk port (the
276 default) and tap0 as an access port for VLAN 9:
279 ovs-vsctl add-port br0 eth0
280 ovs-vsctl add-port br0 tap0 tag=9
282 If you want to configure an already added port as an access port,
283 use "ovs-vsctl set", e.g.:
285 ovs-vsctl set port tap0 tag=9
287 Q: How do I configure a port as a SPAN port, that is, enable mirroring
288 of all traffic to that port?
290 A: The following commands configure br0 with eth0 and tap0 as trunk
291 ports. All traffic coming in or going out on eth0 or tap0 is also
292 mirrored to tap1; any traffic arriving on tap1 is dropped:
295 ovs-vsctl add-port br0 eth0
296 ovs-vsctl add-port br0 tap0
297 ovs-vsctl add-port br0 tap1 \
298 -- --id=@p get port tap1 \
299 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
300 -- set bridge br0 mirrors=@m
302 To later disable mirroring, run:
304 ovs-vsctl clear bridge br0 mirrors
306 Q: Does Open vSwitch support configuring a port in promiscuous mode?
308 A: Yes. How you configure it depends on what you mean by "promiscuous
311 - Conventionally, "promiscuous mode" is a feature of a network
312 interface card. Ordinarily, a NIC passes to the CPU only the
313 packets actually destined to its host machine. It discards
314 the rest to avoid wasting memory and CPU cycles. When
315 promiscuous mode is enabled, however, it passes every packet
316 to the CPU. On an old-style shared-media or hub-based
317 network, this allows the host to spy on all packets on the
318 network. But in the switched networks that are almost
319 everywhere these days, promiscuous mode doesn't have much
320 effect, because few packets not destined to a host are
321 delivered to the host's NIC.
323 This form of promiscuous mode is configured in the guest OS of
324 the VMs on your bridge, e.g. with "ifconfig".
326 - The VMware vSwitch uses a different definition of "promiscuous
327 mode". When you configure promiscuous mode on a VMware vNIC,
328 the vSwitch sends a copy of every packet received by the
329 vSwitch to that vNIC. That has a much bigger effect than just
330 enabling promiscuous mode in a guest OS. Rather than getting
331 a few stray packets for which the switch does not yet know the
332 correct destination, the vNIC gets every packet. The effect
333 is similar to replacing the vSwitch by a virtual hub.
335 This "promiscuous mode" is what switches normally call "port
336 mirroring" or "SPAN". For information on how to configure
337 SPAN, see "How do I configure a port as a SPAN port, that is,
338 enable mirroring of all traffic to that port?"
340 Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable
341 mirroring of all traffic to that VLAN?
343 A: The following commands configure br0 with eth0 as a trunk port and
344 tap0 as an access port for VLAN 10. All traffic coming in or going
345 out on tap0, as well as traffic coming in or going out on eth0 in
346 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
347 VLAN 10, in cases where one is present, is dropped as part of
351 ovs-vsctl add-port br0 eth0
352 ovs-vsctl add-port br0 tap0 tag=10
354 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
356 -- set bridge br0 mirrors=@m
358 To later disable mirroring, run:
360 ovs-vsctl clear bridge br0 mirrors
362 Mirroring to a VLAN can disrupt a network that contains unmanaged
363 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
364 GRE tunnel has fewer caveats than mirroring to a VLAN and should
365 generally be preferred.
367 Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
369 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
370 of the specified output-vlan. This loss of information may make
371 the mirrored traffic too hard to interpret.
373 To mirror multiple VLANs, use the commands above, but specify a
374 comma-separated list of VLANs as the value for select-vlan. To
375 mirror every VLAN, use the commands above, but omit select-vlan and
378 When a packet arrives on a VLAN that is used as a mirror output
379 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
380 floods the packet across all the ports for which the mirror output
381 VLAN is configured. (If an OpenFlow controller is in use, then it
382 can override this behavior through the flow table.) If OVS is used
383 as an intermediate switch, rather than an edge switch, this ensures
384 that the RSPAN traffic is distributed through the network.
386 Mirroring to a VLAN can disrupt a network that contains unmanaged
387 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
388 GRE tunnel has fewer caveats than mirroring to a VLAN and should
389 generally be preferred.
391 Q: How do I configure mirroring of all traffic to a GRE tunnel?
393 A: The following commands configure br0 with eth0 and tap0 as trunk
394 ports. All traffic coming in or going out on eth0 or tap0 is also
395 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
396 traffic arriving on gre0 is dropped:
399 ovs-vsctl add-port br0 eth0
400 ovs-vsctl add-port br0 tap0
401 ovs-vsctl add-port br0 gre0 \
402 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
403 -- --id=@p get port gre0 \
404 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
405 -- set bridge br0 mirrors=@m
407 To later disable mirroring and destroy the GRE tunnel:
409 ovs-vsctl clear bridge br0 mirrors
410 ovs-vcstl del-port br0 gre0
412 Q: Does Open vSwitch support ERSPAN?
414 A: No. ERSPAN is an undocumented proprietary protocol. As an
415 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
418 Q: How do I connect two bridges?
420 A: First, why do you want to do this? Two connected bridges are not
421 much different from a single bridge, so you might as well just have
422 a single bridge with all your ports on it.
424 If you still want to connect two bridges, you can use a pair of
425 patch ports. The following example creates bridges br0 and br1,
426 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
427 and br1 with a pair of patch ports.
430 ovs-vsctl add-port br0 eth0
431 ovs-vsctl add-port br0 tap0
433 ovs-vsctl add-port br1 tap1
435 -- add-port br0 patch0 \
436 -- set interface patch0 type=patch options:peer=patch1 \
437 -- add-port br1 patch1 \
438 -- set interface patch1 type=patch options:peer=patch0
440 Bridges connected with patch ports are much like a single bridge.
441 For instance, if the example above also added eth1 to br1, and both
442 eth0 and eth1 happened to be connected to the same next-hop switch,
443 then you could loop your network just as you would if you added
444 eth0 and eth1 to the same bridge (see the "Configuration Problems"
445 section below for more information).
447 If you are using Open vSwitch 1.9 or an earlier version, then you
448 need to be using the kernel module bundled with Open vSwitch rather
449 than the one that is integrated into Linux 3.3 and later, because
450 Open vSwitch 1.9 and earlier versions need kernel support for patch
451 ports. This also means that in Open vSwitch 1.9 and earlier, patch
452 ports will not work with the userspace datapath, only with the
455 Q: How do I configure a bridge without an OpenFlow local port?
456 (Local port in the sense of OFPP_LOCAL)
458 A: Open vSwitch does not support such a configuration.
459 Bridges always have their local ports.
462 Implementation Details
463 ----------------------
465 Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
467 A: Open vSwitch uses different kinds of flows for different purposes:
469 - OpenFlow flows are the most important kind of flow. OpenFlow
470 controllers use these flows to define a switch's policy.
471 OpenFlow flows support wildcards, priorities, and multiple
474 When in-band control is in use, Open vSwitch sets up a few
475 "hidden" flows, with priority higher than a controller or the
476 user can configure, that are not visible via OpenFlow. (See
477 the "Controller" section of the FAQ for more information
480 - The Open vSwitch software switch implementation uses a second
481 kind of flow internally. These flows, called "datapath" or
482 "kernel" flows, do not support priorities and comprise only a
483 single table, which makes them suitable for caching. (Like
484 OpenFlow flows, datapath flows do support wildcarding, in Open
485 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
486 also support different actions and number ports differently.
488 Datapath flows are an implementation detail that is subject to
489 change in future versions of Open vSwitch. Even with the
490 current version of Open vSwitch, hardware switch
491 implementations do not necessarily use this architecture.
493 Users and controllers directly control only the OpenFlow flow
494 table. Open vSwitch manages the datapath flow table itself, so
495 users should not normally be concerned with it.
497 Q: Why are there so many different ways to dump flows?
499 A: Open vSwitch has two kinds of flows (see the previous question), so
500 it has commands with different purposes for dumping each kind of
503 - "ovs-ofctl dump-flows <br>" dumps OpenFlow flows, excluding
504 hidden flows. This is the most commonly useful form of flow
505 dump. (Unlike the other commands, this should work with any
506 OpenFlow switch, not just Open vSwitch.)
508 - "ovs-appctl bridge/dump-flows <br>" dumps OpenFlow flows,
509 including hidden flows. This is occasionally useful for
510 troubleshooting suspected issues with in-band control.
512 - "ovs-dpctl dump-flows [dp]" dumps the datapath flow table
513 entries for a Linux kernel-based datapath. In Open vSwitch
514 1.10 and later, ovs-vswitchd merges multiple switches into a
515 single datapath, so it will show all the flows on all your
516 kernel-based switches. This command can occasionally be
517 useful for debugging.
519 - "ovs-appctl dpif/dump-flows <br>", new in Open vSwitch 1.10,
520 dumps datapath flows for only the specified bridge, regardless
523 Q: How does multicast snooping works with VLANs?
525 A: Open vSwitch maintains snooping tables for each VLAN.
531 Q: I just upgraded and I see a performance drop. Why?
533 A: The OVS kernel datapath may have been updated to a newer version than
534 the OVS userspace components. Sometimes new versions of OVS kernel
535 module add functionality that is backwards compatible with older
536 userspace components but may cause a drop in performance with them.
537 Especially, if a kernel module from OVS 2.1 or newer is paired with
538 OVS userspace 1.10 or older, there will be a performance drop for
541 Updating the OVS userspace components to the latest released
542 version should fix the performance degradation.
544 To get the best possible performance and functionality, it is
545 recommended to pair the same versions of the kernel module and OVS
549 Configuration Problems
550 ----------------------
552 Q: I created a bridge and added my Ethernet port to it, using commands
556 ovs-vsctl add-port br0 eth0
558 and as soon as I ran the "add-port" command I lost all connectivity
561 A: A physical Ethernet device that is part of an Open vSwitch bridge
562 should not have an IP address. If one does, then that IP address
563 will not be fully functional.
565 You can restore functionality by moving the IP address to an Open
566 vSwitch "internal" device, such as the network device named after
567 the bridge itself. For example, assuming that eth0's IP address is
568 192.168.128.5, you could run the commands below to fix up the
571 ifconfig eth0 0.0.0.0
572 ifconfig br0 192.168.128.5
574 (If your only connection to the machine running OVS is through the
575 IP address in question, then you would want to run all of these
576 commands on a single command line, or put them into a script.) If
577 there were any additional routes assigned to eth0, then you would
578 also want to use commands to adjust these routes to go through br0.
580 If you use DHCP to obtain an IP address, then you should kill the
581 DHCP client that was listening on the physical Ethernet interface
582 (e.g. eth0) and start one listening on the internal interface
583 (e.g. br0). You might still need to manually clear the IP address
584 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
586 There is no compelling reason why Open vSwitch must work this way.
587 However, this is the way that the Linux kernel bridge module has
588 always worked, so it's a model that those accustomed to Linux
589 bridging are already used to. Also, the model that most people
590 expect is not implementable without kernel changes on all the
591 versions of Linux that Open vSwitch supports.
593 By the way, this issue is not specific to physical Ethernet
594 devices. It applies to all network devices except Open vSwitch
597 Q: I created a bridge and added a couple of Ethernet ports to it,
598 using commands like these:
601 ovs-vsctl add-port br0 eth0
602 ovs-vsctl add-port br0 eth1
604 and now my network seems to have melted: connectivity is unreliable
605 (even connectivity that doesn't go through Open vSwitch), all the
606 LEDs on my physical switches are blinking, wireshark shows
607 duplicated packets, and CPU usage is very high.
609 A: More than likely, you've looped your network. Probably, eth0 and
610 eth1 are connected to the same physical Ethernet switch. This
611 yields a scenario where OVS receives a broadcast packet on eth0 and
612 sends it out on eth1, then the physical switch connected to eth1
613 sends the packet back on eth0, and so on forever. More complicated
614 scenarios, involving a loop through multiple switches, are possible
617 The solution depends on what you are trying to do:
619 - If you added eth0 and eth1 to get higher bandwidth or higher
620 reliability between OVS and your physical Ethernet switch,
621 use a bond. The following commands create br0 and then add
622 eth0 and eth1 as a bond:
625 ovs-vsctl add-bond br0 bond0 eth0 eth1
627 Bonds have tons of configuration options. Please read the
628 documentation on the Port table in ovs-vswitchd.conf.db(5)
631 - Perhaps you don't actually need eth0 and eth1 to be on the
632 same bridge. For example, if you simply want to be able to
633 connect each of them to virtual machines, then you can put
634 each of them on a bridge of its own:
637 ovs-vsctl add-port br0 eth0
640 ovs-vsctl add-port br1 eth1
642 and then connect VMs to br0 and br1. (A potential
643 disadvantage is that traffic cannot directly pass between br0
644 and br1. Instead, it will go out eth0 and come back in eth1,
647 - If you have a redundant or complex network topology and you
648 want to prevent loops, turn on spanning tree protocol (STP).
649 The following commands create br0, enable STP, and add eth0
650 and eth1 to the bridge. The order is important because you
651 don't want have to have a loop in your network even
655 ovs-vsctl set bridge br0 stp_enable=true
656 ovs-vsctl add-port br0 eth0
657 ovs-vsctl add-port br0 eth1
659 The Open vSwitch implementation of STP is not well tested.
660 Please report any bugs you observe, but if you'd rather avoid
661 acting as a beta tester then another option might be your
664 Q: I can't seem to use Open vSwitch in a wireless network.
666 A: Wireless base stations generally only allow packets with the source
667 MAC address of NIC that completed the initial handshake.
668 Therefore, without MAC rewriting, only a single device can
669 communicate over a single wireless link.
671 This isn't specific to Open vSwitch, it's enforced by the access
672 point, so the same problems will show up with the Linux bridge or
673 any other way to do bridging.
675 Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
677 A: PPP most commonly carries IP packets, but Open vSwitch works only
678 with Ethernet frames. The correct way to interface PPP to an
679 Ethernet network is usually to use routing instead of switching.
681 Q: Is there any documentation on the database tables and fields?
683 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
685 Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
686 I only see a datapath called "ovs-system". How can I see datapath
687 information about a particular bridge?
689 A: In version 1.9.0, OVS switched to using a single datapath that is
690 shared by all bridges of that type. The "ovs-appctl dpif/*"
691 commands provide similar functionality that is scoped by the bridge.
693 Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
694 see the port in the datapath?
696 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
697 cannot be loaded at the same time. It is likely that on your system the
698 Linux GRE module is already loaded and blocking OVS (to confirm, check
699 dmesg for errors regarding GRE registration). To fix this, unload all
700 GRE modules that appear in lsmod as well as the OVS kernel module. You
701 can then reload the OVS module following the directions in INSTALL,
702 which will ensure that dependencies are satisfied.
704 Q: Open vSwitch does not seem to obey my packet filter rules.
706 A: It depends on mechanisms and configurations you want to use.
708 You cannot usefully use typical packet filters, like iptables, on
709 physical Ethernet ports that you add to an Open vSwitch bridge.
710 This is because Open vSwitch captures packets from the interface at
711 a layer lower below where typical packet-filter implementations
712 install their hooks. (This actually applies to any interface of
713 type "system" that you might add to an Open vSwitch bridge.)
715 You can usefully use typical packet filters on Open vSwitch
716 internal ports as they are mostly ordinary interfaces from the point
717 of view of packet filters.
719 For example, suppose you create a bridge br0 and add Ethernet port
720 eth0 to it. Then you can usefully add iptables rules to affect the
721 internal interface br0, but not the physical interface eth0. (br0
722 is also where you would add an IP address, as discussed elsewhere
725 For simple filtering rules, it might be possible to achieve similar
726 results by installing appropriate OpenFlow flows instead.
728 If the use of a particular packet filter setup is essential, Open
729 vSwitch might not be the best choice for you. On Linux, you might
730 want to consider using the Linux Bridge. (This is the only choice if
731 you want to use ebtables rules.) On NetBSD, you might want to
732 consider using the bridge(4) with BRIDGE_IPF option.
734 Q: It seems that Open vSwitch does nothing when I removed a port and
735 then immediately put it back. For example, consider that p1 is
736 a port of type=internal:
738 ovs-vsctl del-port br0 p1 -- \
740 set interface p1 type=internal
742 A: It's an expected behaviour.
744 If del-port and add-port happen in a single OVSDB transaction as
745 your example, Open vSwitch always "skips" the intermediate steps.
746 Even if they are done in multiple transactions, it's still allowed
747 for Open vSwitch to skip the intermediate steps and just implement
748 the overall effect. In both cases, your example would be turned
751 If you want to make Open vSwitch actually destroy and then re-create
752 the port for some side effects like resetting kernel setting for the
753 corresponding interface, you need to separate operations into multiple
754 OVSDB transactions and ensure that at least the first one does not have
755 --no-wait. In the following example, the first ovs-vsctl will block
756 until Open vSwitch reloads the new configuration and removes the port:
758 ovs-vsctl del-port br0 p1
759 ovs-vsctl add-port br0 p1 -- \
760 set interface p1 type=internal
762 Quality of Service (QoS)
763 ------------------------
765 Q: How do I configure Quality of Service (QoS)?
767 A: Suppose that you want to set up bridge br0 connected to physical
768 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
769 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
770 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
771 could configure the bridge this way:
775 add-port br0 eth0 -- \
776 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
777 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
778 set port eth0 qos=@newqos -- \
779 --id=@newqos create qos type=linux-htb \
780 other-config:max-rate=1000000000 \
781 queues:123=@vif10queue \
782 queues:234=@vif20queue -- \
783 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
784 --id=@vif20queue create queue other-config:max-rate=20000000
786 At this point, bridge br0 is configured with the ports and eth0 is
787 configured with the queues that you need for QoS, but nothing is
788 actually directing packets from vif1.0 or vif2.0 to the queues that
789 we have set up for them. That means that all of the packets to
790 eth0 are going to the "default queue", which is not what we want.
792 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
793 queues reserved for them:
795 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
796 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
798 Each of the above flows matches on the input port, sets up the
799 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
800 executes the "normal" action, which performs the same switching
801 that Open vSwitch would have done without any OpenFlow flows being
802 present. (We know that vif1.0 and vif2.0 have OpenFlow port
803 numbers 5 and 6, respectively, because we set their ofport_request
804 columns above. If we had not done that, then we would have needed
805 to find out their port numbers before setting up these flows.)
807 Now traffic going from vif1.0 or vif2.0 to eth0 should be
810 By the way, if you delete the bridge created by the above commands,
815 then that will leave one unreferenced QoS record and two
816 unreferenced Queue records in the Open vSwich database. One way to
817 clear them out, assuming you don't have other QoS or Queue records
818 that you want to keep, is:
820 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
822 If you do want to keep some QoS or Queue records, or the Open
823 vSwitch you are using is older than version 1.8 (which added the
824 --all option), then you will have to destroy QoS and Queue records
827 Q: I configured Quality of Service (QoS) in my OpenFlow network by
828 adding records to the QoS and Queue table, but the results aren't
831 A: Did you install OpenFlow flows that use your queues? This is the
832 primary way to tell Open vSwitch which queues you want to use. If
833 you don't do this, then the default queue will be used, which will
834 probably not have the effect you want.
836 Refer to the previous question for an example.
838 Q: I'd like to take advantage of some QoS feature that Open vSwitch
839 doesn't yet support. How do I do that?
841 A: Open vSwitch does not implement QoS itself. Instead, it can
842 configure some, but not all, of the QoS features built into the
843 Linux kernel. If you need some QoS feature that OVS cannot
844 configure itself, then the first step is to figure out whether
845 Linux QoS supports that feature. If it does, then you can submit a
846 patch to support Open vSwitch configuration for that feature, or
847 you can use "tc" directly to configure the feature in Linux. (If
848 Linux QoS doesn't support the feature you want, then first you have
849 to add that support to Linux.)
851 Q: I configured QoS, correctly, but my measurements show that it isn't
852 working as well as I expect.
854 A: With the Linux kernel, the Open vSwitch implementation of QoS has
857 - Open vSwitch configures a subset of Linux kernel QoS
858 features, according to what is in OVSDB. It is possible that
859 this code has bugs. If you believe that this is so, then you
860 can configure the Linux traffic control (QoS) stack directly
861 with the "tc" program. If you get better results that way,
862 you can send a detailed bug report to bugs@openvswitch.org.
864 It is certain that Open vSwitch cannot configure every Linux
865 kernel QoS feature. If you need some feature that OVS cannot
866 configure, then you can also use "tc" directly (or add that
869 - The Open vSwitch implementation of OpenFlow allows flows to
870 be directed to particular queues. This is pretty simple and
871 unlikely to have serious bugs at this point.
873 However, most problems with QoS on Linux are not bugs in Open
874 vSwitch at all. They tend to be either configuration errors
875 (please see the earlier questions in this section) or issues with
876 the traffic control (QoS) stack in Linux. The Open vSwitch
877 developers are not experts on Linux traffic control. We suggest
878 that, if you believe you are encountering a problem with Linux
879 traffic control, that you consult the tc manpages (e.g. tc(8),
880 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
881 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
883 Q: Does Open vSwitch support OpenFlow meters?
885 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
886 OpenFlow meters. There is no implementation of meters in the Open
887 vSwitch software switch (neither the kernel-based nor userspace
896 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
897 partition a single switch into multiple switches. Suppose, for
898 example, that you have two groups of machines, group A and group B.
899 You want the machines in group A to be able to talk to each other,
900 and you want the machine in group B to be able to talk to each
901 other, but you don't want the machines in group A to be able to
902 talk to the machines in group B. You can do this with two
903 switches, by plugging the machines in group A into one switch and
904 the machines in group B into the other switch.
906 If you only have one switch, then you can use VLANs to do the same
907 thing, by configuring the ports for machines in group A as VLAN
908 "access ports" for one VLAN and the ports for group B as "access
909 ports" for a different VLAN. The switch will only forward packets
910 between ports that are assigned to the same VLAN, so this
911 effectively subdivides your single switch into two independent
912 switches, one for each group of machines.
914 So far we haven't said anything about VLAN headers. With access
915 ports, like we've described so far, no VLAN header is present in
916 the Ethernet frame. This means that the machines (or switches)
917 connected to access ports need not be aware that VLANs are
918 involved, just like in the case where we use two different physical
921 Now suppose that you have a whole bunch of switches in your
922 network, instead of just one, and that some machines in group A are
923 connected directly to both switches 1 and 2. To allow these
924 machines to talk to each other, you could add an access port for
925 group A's VLAN to switch 1 and another to switch 2, and then
926 connect an Ethernet cable between those ports. That works fine,
927 but it doesn't scale well as the number of switches and the number
928 of VLANs increases, because you use up a lot of valuable switch
929 ports just connecting together your VLANs.
931 This is where VLAN headers come in. Instead of using one cable and
932 two ports per VLAN to connect a pair of switches, we configure a
933 port on each switch as a VLAN "trunk port". Packets sent and
934 received on a trunk port carry a VLAN header that says what VLAN
935 the packet belongs to, so that only two ports total are required to
936 connect the switches, regardless of the number of VLANs in use.
937 Normally, only switches (either physical or virtual) are connected
938 to a trunk port, not individual hosts, because individual hosts
939 don't expect to see a VLAN header in the traffic that they receive.
941 None of the above discussion says anything about particular VLAN
942 numbers. This is because VLAN numbers are completely arbitrary.
943 One must only ensure that a given VLAN is numbered consistently
944 throughout a network and that different VLANs are given different
945 numbers. (That said, VLAN 0 is usually synonymous with a packet
946 that has no VLAN header, and VLAN 4095 is reserved.)
950 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
951 bugs. If you are having problems with VLANs that you suspect to be
952 driver related, then you have several options:
954 - Upgrade to Linux 3.3 or later.
956 - Build and install a fixed version of the particular driver
957 that is causing trouble, if one is available.
959 - Use a NIC whose driver does not have VLAN problems.
961 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
962 that works around bugs in kernel drivers. To enable VLAN
963 splinters on interface eth0, use the command:
965 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
967 For VLAN splinters to be effective, Open vSwitch must know
968 which VLANs are in use. See the "VLAN splinters" section in
969 the Interface table in ovs-vswitchd.conf.db(5) for details on
970 how Open vSwitch infers in-use VLANs.
972 VLAN splinters increase memory use and reduce performance, so
973 use them only if needed.
975 - Apply the "vlan workaround" patch from the XenServer kernel
976 patch queue, build Open vSwitch against this patched kernel,
977 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
978 workaround for each interface whose driver is buggy.
980 (This is a nontrivial exercise, so this option is included
981 only for completeness.)
983 It is not always easy to tell whether a Linux kernel driver has
984 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
985 can help you test. See their manpages for details. Of the two
986 utilities, ovs-test(8) is newer and more thorough, but
987 ovs-vlan-test(8) may be easier to use.
989 Q: VLANs still don't work. I've tested the driver so I know that it's OK.
991 A: Do you have VLANs enabled on the physical switch that OVS is
992 attached to? Make sure that the port is configured to trunk the
993 VLAN or VLANs that you are using with OVS.
995 Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
996 and to its destination host, but OVS seems to drop incoming return
999 A: It's possible that you have the VLAN configured on your physical
1000 switch as the "native" VLAN. In this mode, the switch treats
1001 incoming packets either tagged with the native VLAN or untagged as
1002 part of the native VLAN. It may also send outgoing packets in the
1003 native VLAN without a VLAN tag.
1005 If this is the case, you have two choices:
1007 - Change the physical switch port configuration to tag packets
1008 it forwards to OVS with the native VLAN instead of forwarding
1011 - Change the OVS configuration for the physical port to a
1012 native VLAN mode. For example, the following sets up a
1013 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1015 ovs-vsctl add-br br0
1016 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1018 In this situation, "native-untagged" mode will probably work
1019 equally well. Refer to the documentation for the Port table
1020 in ovs-vswitchd.conf.db(5) for more information.
1022 Q: I added a pair of VMs on different VLANs, like this:
1024 ovs-vsctl add-br br0
1025 ovs-vsctl add-port br0 eth0
1026 ovs-vsctl add-port br0 tap0 tag=9
1027 ovs-vsctl add-port br0 tap1 tag=10
1029 but the VMs can't access each other, the external network, or the
1032 A: It is to be expected that the VMs can't access each other. VLANs
1033 are a means to partition a network. When you configured tap0 and
1034 tap1 as access ports for different VLANs, you indicated that they
1035 should be isolated from each other.
1037 As for the external network and the Internet, it seems likely that
1038 the machines you are trying to access are not on VLAN 9 (or 10) and
1039 that the Internet is not available on VLAN 9 (or 10).
1041 Q: I added a pair of VMs on the same VLAN, like this:
1043 ovs-vsctl add-br br0
1044 ovs-vsctl add-port br0 eth0
1045 ovs-vsctl add-port br0 tap0 tag=9
1046 ovs-vsctl add-port br0 tap1 tag=9
1048 The VMs can access each other, but not the external network or the
1051 A: It seems likely that the machines you are trying to access in the
1052 external network are not on VLAN 9 and that the Internet is not
1053 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1054 trunk VLAN on the upstream switch port to which eth0 is connected.
1056 Q: Can I configure an IP address on a VLAN?
1058 A: Yes. Use an "internal port" configured as an access port. For
1059 example, the following configures IP address 192.168.0.7 on VLAN 9.
1060 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1061 they have an 802.1Q header with VLAN 9. Conversely, traffic
1062 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1065 ovs-vsctl add-br br0
1066 ovs-vsctl add-port br0 eth0
1067 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1068 ifconfig vlan9 192.168.0.7
1070 See also the following question.
1072 Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1075 ovs-vsctl add-br br0
1076 ovs-vsctl add-port br0 eth0
1077 ifconfig br0 192.168.0.5
1078 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1079 ifconfig vlan9 192.168.0.9
1081 but other hosts that are only on VLAN 0 can reach the IP address
1082 configured on VLAN 9. What's going on?
1084 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1085 approaches to IP address handling in Internet hosts:
1087 - In the "Strong ES Model", where an ES is a host ("End
1088 System"), an IP address is primarily associated with a
1089 particular interface. The host discards packets that arrive
1090 on interface A if they are destined for an IP address that is
1091 configured on interface B. The host never sends packets from
1092 interface A using a source address configured on interface B.
1094 - In the "Weak ES Model", an IP address is primarily associated
1095 with a host. The host accepts packets that arrive on any
1096 interface if they are destined for any of the host's IP
1097 addresses, even if the address is configured on some
1098 interface other than the one on which it arrived. The host
1099 does not restrict itself to sending packets from an IP
1100 address associated with the originating interface.
1102 Linux uses the weak ES model. That means that when packets
1103 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1104 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1105 address, even though they were not received on vlan9, the network
1108 To simulate the strong ES model on Linux, one may add iptables rule
1109 to filter packets based on source and destination address and
1110 adjust ARP configuration with sysctls.
1112 BSD uses the strong ES model.
1114 Q: My OpenFlow controller doesn't see the VLANs that I expect.
1116 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1117 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1118 implementation of the OpenFlow "normal switching" action. By
1119 default, when Open vSwitch isn't connected to a controller and
1120 nothing has been manually configured in the flow table, all traffic
1121 goes through the "normal switching" action. But, if you set up
1122 OpenFlow flows on your own, through a controller or using ovs-ofctl
1123 or through other means, then you have to implement VLAN handling
1126 You can use "normal switching" as a component of your OpenFlow
1127 actions, e.g. by putting "normal" into the lists of actions on
1128 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1129 controller. In situations where this is not suitable, you can
1130 implement VLAN handling yourself, e.g.:
1132 - If a packet comes in on an access port, and the flow table
1133 needs to send it out on a trunk port, then the flow can add
1134 the appropriate VLAN tag with the "mod_vlan_vid" action.
1136 - If a packet comes in on a trunk port, and the flow table
1137 needs to send it out on an access port, then the flow can
1138 strip the VLAN tag with the "strip_vlan" action.
1140 Q: I configured ports on a bridge as access ports with different VLAN
1143 ovs-vsctl add-br br0
1144 ovs-vsctl set-controller br0 tcp:192.168.0.10:6633
1145 ovs-vsctl add-port br0 eth0
1146 ovs-vsctl add-port br0 tap0 tag=9
1147 ovs-vsctl add-port br0 tap1 tag=10
1149 but the VMs running behind tap0 and tap1 can still communicate,
1150 that is, they are not isolated from each other even though they are
1153 A: Do you have a controller configured on br0 (as the commands above
1154 do)? If so, then this is a variant on the previous question, "My
1155 OpenFlow controller doesn't see the VLANs that I expect," and you
1156 can refer to the answer there for more information.
1158 Q: How MAC learning works with VLANs?
1160 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1161 OFPP_NORMAL action. I.e. it logically has separate learning tables
1170 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1171 to solve the scaling challenges of VLAN networks in a multi-tenant
1172 environment. VXLAN is an overlay network which transports an L2 network
1173 over an existing L3 network. For more information on VXLAN, please see
1174 the IETF draft available here:
1176 http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-03
1178 Q: How much of the VXLAN protocol does Open vSwitch currently support?
1180 A: Open vSwitch currently supports the framing format for packets on the
1181 wire. There is currently no support for the multicast aspects of VXLAN.
1182 To get around the lack of multicast support, it is possible to
1183 pre-provision MAC to IP address mappings either manually or from a
1186 Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1189 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1190 is 4789. However, it is possible to configure the destination UDP port
1191 manually on a per-VXLAN tunnel basis. An example of this configuration is
1194 ovs-vsctl add-br br0
1195 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1196 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1197 options:dst_port=8472
1200 Using OpenFlow (Manually or Via Controller)
1201 -------------------------------------------
1203 Q: What versions of OpenFlow does Open vSwitch support?
1205 A: The following table lists the versions of OpenFlow supported by
1206 each version of Open vSwitch:
1208 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5
1209 =============== ===== ===== ===== ===== ===== =====
1210 1.9 and earlier yes --- --- --- --- ---
1211 1.10 yes --- [*] [*] --- ---
1212 1.11 yes --- [*] [*] --- ---
1213 2.0 yes [*] [*] [*] --- ---
1214 2.1 yes [*] [*] [*] --- ---
1215 2.2 yes [*] [*] [*] [%] [*]
1216 2.3 yes yes yes yes [*] [*]
1218 [*] Supported, with one or more missing features.
1219 [%] Experimental, unsafe implementation.
1221 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1222 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1223 1.2, and 1.3 must be enabled manually in ovs-vswitchd. OpenFlow
1224 1.4 and 1.5 are also supported, with missing features, in Open
1225 vSwitch 2.3 and later, but not enabled by default. In any case,
1226 the user may override the default:
1228 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1230 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1232 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1234 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1236 - To enable only OpenFlow 1.0 on bridge br0:
1238 ovs-vsctl set bridge br0 protocols=OpenFlow10
1240 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1241 default. Use the -O option to enable support for later versions of
1242 OpenFlow in ovs-ofctl. For example:
1244 ovs-ofctl -O OpenFlow13 dump-flows br0
1246 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1247 1.4 that could cause crashes. We don't recommend enabling it.)
1249 OPENFLOW-1.1+ in the Open vSwitch source tree tracks support for
1250 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1251 1.5 is solidly implemented, Open vSwitch will enable those version
1252 by default. Also, the OpenFlow 1.5 specification is still under
1253 development and thus subject to change.
1255 Q: Does Open vSwitch support MPLS?
1257 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1258 these versions can match on MPLS Ethernet types, but they cannot
1259 match, push, or pop MPLS labels, nor can they look past MPLS labels
1260 into the encapsulated packet.
1262 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1263 for MPLS. With the userspace datapath only, these versions can
1264 match, push, or pop a single MPLS label, but they still cannot look
1265 past MPLS labels (even after popping them) into the encapsulated
1266 packet. Kernel datapath support is unchanged from earlier
1269 Open vSwitch version 2.3 can match, push, or pop up to 3 MPLS
1270 labels. Looking past MPLS labels into the encapsulated packet will
1271 still be unsupported. Both userspace and kernel datapaths will be
1272 supported, but MPLS processing always happens in userspace either
1273 way, so kernel datapath performance will be disappointing.
1275 Open vSwitch version 2.4 will have kernel support for MPLS,
1276 yielding improved performance.
1278 Q: I'm getting "error type 45250 code 0". What's that?
1280 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1281 vSwitch uses this extension when it must report an error to an
1282 OpenFlow controller but no standard OpenFlow error code is
1285 Open vSwitch logs the errors that it sends to controllers, so the
1286 easiest thing to do is probably to look at the ovs-vswitchd log to
1287 find out what the error was.
1289 If you want to dissect the extended error message yourself, the
1290 format is documented in include/openflow/nicira-ext.h in the Open
1291 vSwitch source distribution. The extended error codes are
1292 documented in lib/ofp-errors.h.
1294 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1295 doesn't actually appear through the OpenFlow connection, even
1296 though I know that it's going through.
1297 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1298 to apply to certain traffic, especially traffic between OVS and
1299 the controller itself.
1301 A: By default, Open vSwitch assumes that OpenFlow controllers are
1302 connected "in-band", that is, that the controllers are actually
1303 part of the network that is being controlled. In in-band mode,
1304 Open vSwitch sets up special "hidden" flows to make sure that
1305 traffic can make it back and forth between OVS and the controllers.
1306 These hidden flows are higher priority than any flows that can be
1307 set up through OpenFlow, and they are not visible through normal
1308 OpenFlow flow table dumps.
1310 Usually, the hidden flows are desirable and helpful, but
1311 occasionally they can cause unexpected behavior. You can view the
1312 full OpenFlow flow table, including hidden flows, on bridge br0
1315 ovs-appctl bridge/dump-flows br0
1317 to help you debug. The hidden flows are those with priorities
1318 greater than 65535 (the maximum priority that can be set with
1321 The DESIGN file at the top level of the Open vSwitch source
1322 distribution describes the in-band model in detail.
1324 If your controllers are not actually in-band (e.g. they are on
1325 localhost via 127.0.0.1, or on a separate network), then you should
1326 configure your controllers in "out-of-band" mode. If you have one
1327 controller on bridge br0, then you can configure out-of-band mode
1330 ovs-vsctl set controller br0 connection-mode=out-of-band
1332 Q: I configured all my controllers for out-of-band control mode but
1333 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1335 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1336 set-manager"). By default, Open vSwitch assumes that managers need
1337 in-band rules set up on every bridge. You can disable these rules
1340 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1342 This actually disables in-band control entirely for the bridge, as
1343 if all the bridge's controllers were configured for out-of-band
1346 Q: My OpenFlow controller doesn't see the VLANs that I expect.
1348 A: See answer under "VLANs", above.
1350 Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1351 but I got a funny message like this:
1353 ofp_util|INFO|normalization changed ofp_match, details:
1354 ofp_util|INFO| pre: nw_dst=192.168.0.1
1357 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1358 match had disappeared, so that the flow ends up matching every
1361 A: The term "normalization" in the log message means that a flow
1362 cannot match on an L3 field without saying what L3 protocol is in
1363 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1364 so the L3 field match was dropped.
1366 In this case, the L3 protocol could be IP or ARP. A correct
1367 command for each possibility is, respectively:
1369 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1373 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1375 Similarly, a flow cannot match on an L4 field without saying what
1376 L4 protocol is in use. For example, the flow match "tp_src=1234"
1377 is, by itself, meaningless and will be ignored. Instead, to match
1378 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1379 source port 1234, write "udp,tp_src=1234".
1381 Q: How can I figure out the OpenFlow port number for a given port?
1383 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1384 respond with an OFPT_FEATURES_REPLY that, among other information,
1385 includes a mapping between OpenFlow port names and numbers. From a
1386 command prompt, "ovs-ofctl show br0" makes such a request and
1387 prints the response for switch br0.
1389 The Interface table in the Open vSwitch database also maps OpenFlow
1390 port names to numbers. To print the OpenFlow port number
1391 associated with interface eth0, run:
1393 ovs-vsctl get Interface eth0 ofport
1395 You can print the entire mapping with:
1397 ovs-vsctl -- --columns=name,ofport list Interface
1399 but the output mixes together interfaces from all bridges in the
1400 database, so it may be confusing if more than one bridge exists.
1402 In the Open vSwitch database, ofport value -1 means that the
1403 interface could not be created due to an error. (The Open vSwitch
1404 log should indicate the reason.) ofport value [] (the empty set)
1405 means that the interface hasn't been created yet. The latter is
1406 normally an intermittent condition (unless ovs-vswitchd is not
1409 Q: I added some flows with my controller or with ovs-ofctl, but when I
1410 run "ovs-dpctl dump-flows" I don't see them.
1412 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1413 won't display the information that you want. You want to use
1414 "ovs-ofctl dump-flows" instead.
1416 Q: It looks like each of the interfaces in my bonded port shows up
1417 as an individual OpenFlow port. Is that right?
1419 A: Yes, Open vSwitch makes individual bond interfaces visible as
1420 OpenFlow ports, rather than the bond as a whole. The interfaces
1421 are treated together as a bond for only a few purposes:
1423 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1424 controller is not configured, this happens implicitly to
1427 - Mirrors configured for output to a bonded port.
1429 It would make a lot of sense for Open vSwitch to present a bond as
1430 a single OpenFlow port. If you want to contribute an
1431 implementation of such a feature, please bring it up on the Open
1432 vSwitch development mailing list at dev@openvswitch.org.
1434 Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1435 multiple hosts, and other components. The behavior isn't what I
1438 A: To debug network behavior problems, trace the path of a packet,
1439 hop-by-hop, from its origin in one host to a remote host. If
1440 that's correct, then trace the path of the response packet back to
1443 Usually a simple ICMP echo request and reply ("ping") packet is
1444 good enough. Start by initiating an ongoing "ping" from the origin
1445 host to a remote host. If you are tracking down a connectivity
1446 problem, the "ping" will not display any successful output, but
1447 packets are still being sent. (In this case the packets being sent
1448 are likely ARP rather than ICMP.)
1450 Tools available for tracing include the following:
1452 - "tcpdump" and "wireshark" for observing hops across network
1453 devices, such as Open vSwitch internal devices and physical
1456 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1457 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1458 These tools allow one to observe the actions being taken on
1459 packets in ongoing flows.
1461 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1462 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1463 documentation, and "Why are there so many different ways to
1464 dump flows?" above for some background.
1466 - "ovs-appctl ofproto/trace" to observe the logic behind how
1467 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1468 documentation. You can out more details about a given flow
1469 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1470 a flow from the output into an "ovs-appctl ofproto/trace"
1473 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1474 observe what goes on at these physical hops.
1476 Starting at the origin of a given packet, observe the packet at
1477 each hop in turn. For example, in one plausible scenario, you
1480 1. "tcpdump" the "eth" interface through which an ARP egresses
1481 a VM, from inside the VM.
1483 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1484 ingresses the host machine.
1486 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1487 the host interface through which the ARP egresses the
1488 physical machine. You may need to use "ovs-dpctl show" to
1489 interpret the port numbers. If the output seems surprising,
1490 you can use "ovs-appctl ofproto/trace" to observe details of
1491 how ovs-vswitchd determined the actions in the "ovs-dpctl
1494 4. "tcpdump" the "eth" interface through which the ARP egresses
1495 the physical machine.
1497 5. "tcpdump" the "eth" interface through which the ARP
1498 ingresses the physical machine, at the remote host that
1501 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1502 remote host that receives the ARP and observe the VM "vif"
1503 or "tap" interface to which the flow is directed. Again,
1504 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1506 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1509 8. "tcpdump" the "eth" interface through which the ARP
1510 ingresses a VM, from inside the VM.
1512 It is likely that during one of these steps you will figure out the
1513 problem. If not, then follow the ARP reply back to the origin, in
1516 Q: How do I make a flow drop packets?
1518 A: To drop a packet is to receive it without forwarding it. OpenFlow
1519 explicitly specifies forwarding actions. Thus, a flow with an
1520 empty set of actions does not forward packets anywhere, causing
1521 them to be dropped. You can specify an empty set of actions with
1522 "actions=" on the ovs-ofctl command line. For example:
1524 ovs-ofctl add-flow br0 priority=65535,actions=
1526 would cause every packet entering switch br0 to be dropped.
1528 You can write "drop" explicitly if you like. The effect is the
1529 same. Thus, the following command also causes every packet
1530 entering switch br0 to be dropped:
1532 ovs-ofctl add-flow br0 priority=65535,actions=drop
1534 "drop" is not an action, either in OpenFlow or Open vSwitch.
1535 Rather, it is only a way to say that there are no actions.
1537 Q: I added a flow to send packets out the ingress port, like this:
1539 ovs-ofctl add-flow br0 in_port=2,actions=2
1541 but OVS drops the packets instead.
1543 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1544 out its ingress port. The rationale is that dropping these packets
1545 makes it harder to loop the network. Sometimes this behavior can
1546 even be convenient, e.g. it is often the desired behavior in a flow
1547 that forwards a packet to several ports ("floods" the packet).
1549 Sometimes one really needs to send a packet out its ingress port
1550 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1551 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1553 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1555 This also works in some circumstances where the flow doesn't match
1556 on the input port. For example, if you know that your switch has
1557 five ports numbered 2 through 6, then the following will send every
1558 received packet out every port, even its ingress port:
1560 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1564 ovs-ofctl add-flow br0 actions=all,in_port
1566 Sometimes, in complicated flow tables with multiple levels of
1567 "resubmit" actions, a flow needs to output to a particular port
1568 that may or may not be the ingress port. It's difficult to take
1569 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1570 provides, as an OpenFlow extension, the ability to modify the
1571 in_port field. Whatever value is currently in the in_port field is
1572 the port to which outputs will be dropped, as well as the
1573 destination for OFPP_IN_PORT. This means that the following will
1574 reliably output to port 2 or to ports 2 through 6, respectively:
1576 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1577 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1579 If the input port is important, then one may save and restore it on
1582 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1583 load:0->NXM_OF_IN_PORT[],\
1585 pop:NXM_OF_IN_PORT[]
1587 Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1588 on port 2. I set up flows to forward only traffic destined to the
1589 other host and drop other traffic, like this:
1591 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1592 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1593 priority=0,actions=drop
1595 But it doesn't work--I don't get any connectivity when I do this.
1598 A: These flows drop the ARP packets that IP hosts use to establish IP
1599 connectivity over Ethernet. To solve the problem, add flows to
1600 allow ARP to pass between the hosts:
1602 priority=5,in_port=1,arp,actions=2
1603 priority=5,in_port=2,arp,actions=1
1605 This issue can manifest other ways, too. The following flows that
1606 match on Ethernet addresses instead of IP addresses will also drop
1607 ARP packets, because ARP requests are broadcast instead of being
1608 directed to a specific host:
1610 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1611 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1612 priority=0,actions=drop
1614 The solution already described above will also work in this case.
1615 It may be better to add flows to allow all multicast and broadcast
1618 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1619 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1621 Q: My bridge disconnects from my controller on add-port/del-port.
1623 A: Reconfiguring your bridge can change your bridge's datapath-id because
1624 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1625 In that case, Open vSwitch disconnects from controllers because there's
1626 no graceful way to notify controllers about the change of datapath-id.
1628 To avoid the behaviour, you can configure datapath-id manually.
1630 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1636 Q: How do I implement a new OpenFlow message?
1638 A: Add your new message to "enum ofpraw" and "enum ofptype" in
1639 lib/ofp-msgs.h, following the existing pattern. Then recompile and
1640 fix all of the new warnings, implementing new functionality for the
1641 new message as needed. (If you configure with --enable-Werror, as
1642 described in INSTALL, then it is impossible to miss any warnings.)
1644 If you need to add an OpenFlow vendor extension message for a
1645 vendor that doesn't yet have any extension messages, then you will
1646 also need to edit build-aux/extract-ofp-msgs.
1652 bugs@openvswitch.org
1653 http://openvswitch.org/