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 2.6.18 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-controller, ovs-ofctl, ovs-vsctl) that developers
93 can script and extend to provide distributed vswitch capabilities
94 that are closely integrated with their virtualization management
97 Q: Why doesn't Open vSwitch support distribution?
99 A: Open vSwitch is intended to be a useful component for building
100 flexible network infrastructure. There are many different approaches
101 to distribution which balance trade-offs between simplicity,
102 scalability, hardware compatibility, convergence times, logical
103 forwarding model, etc. The goal of Open vSwitch is to be able to
104 support all as a primitive building block rather than choose a
105 particular point in the distributed design space.
107 Q: How can I contribute to the Open vSwitch Community?
109 A: You can start by joining the mailing lists and helping to answer
110 questions. You can also suggest improvements to documentation. If
111 you have a feature or bug you would like to work on, send a mail to
112 one of the mailing lists:
114 http://openvswitch.org/mlists/
120 Q: What does it mean for an Open vSwitch release to be LTS (long-term
123 A: All official releases have been through a comprehensive testing
124 process and are suitable for production use. Planned releases will
125 occur several times a year. If a significant bug is identified in an
126 LTS release, we will provide an updated release that includes the
127 fix. Releases that are not LTS may not be fixed and may just be
128 supplanted by the next major release. The current LTS release is
131 Q: What Linux kernel versions does each Open vSwitch release work with?
133 A: The following table lists the Linux kernel versions against which the
134 given versions of the Open vSwitch kernel module will successfully
135 build. The Linux kernel versions are upstream kernel versions, so
136 Linux kernels modified from the upstream sources may not build in
137 some cases even if they are based on a supported version. This is
138 most notably true of Red Hat Enterprise Linux (RHEL) kernels, which
139 are extensively modified from upstream.
141 Open vSwitch Linux kernel
142 ------------ -------------
152 Open vSwitch userspace should also work with the Linux kernel module
153 built into Linux 3.3 and later.
155 Open vSwitch userspace is not sensitive to the Linux kernel version.
156 It should build against almost any kernel, certainly against 2.6.18
159 Q: What Linux kernel versions does IPFIX flow monitoring work with?
161 A: IPFIX flow monitoring requires the Linux kernel module from Open
162 vSwitch version 1.10.90 or later.
164 Q: Should userspace or kernel be upgraded first to minimize downtime?
166 In general, the Open vSwitch userspace should be used with the
167 kernel version included in the same release or with the version
168 from upstream Linux. However, when upgrading between two releases
169 of Open vSwitch it is best to migrate userspace first to reduce
170 the possbility of incompatibilities.
172 Q: What features are not available in the Open vSwitch kernel datapath
173 that ships as part of the upstream Linux kernel?
175 A: The kernel module in upstream Linux 3.3 and later does not include
176 tunnel virtual ports, that is, interfaces with type "gre",
177 "ipsec_gre", "gre64", "ipsec_gre64", "vxlan", or "lisp". It is
178 possible to create tunnels in Linux and attach them to Open vSwitch
179 as system devices. However, they cannot be dynamically created
180 through the OVSDB protocol or set the tunnel ids as a flow action.
182 Work is in progress in adding tunnel virtual ports to the upstream
183 Linux version of the Open vSwitch kernel module. For now, if you
184 need these features, use the kernel module from the Open vSwitch
185 distribution instead of the upstream Linux kernel module.
187 The upstream kernel module does not include patch ports, but this
188 only matters for Open vSwitch 1.9 and earlier, because Open vSwitch
189 1.10 and later implement patch ports without using this kernel
192 Q: What features are not available when using the userspace datapath?
194 A: Tunnel virtual ports are not supported, as described in the
195 previous answer. It is also not possible to use queue-related
196 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
197 may not be transmitted.
203 Q: I thought Open vSwitch was a virtual Ethernet switch, but the
204 documentation keeps talking about bridges. What's a bridge?
206 A: In networking, the terms "bridge" and "switch" are synonyms. Open
207 vSwitch implements an Ethernet switch, which means that it is also
212 A: See the "VLAN" section below.
218 Q: How do I configure a port as an access port?
220 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
221 the following commands configure br0 with eth0 as a trunk port (the
222 default) and tap0 as an access port for VLAN 9:
225 ovs-vsctl add-port br0 eth0
226 ovs-vsctl add-port br0 tap0 tag=9
228 If you want to configure an already added port as an access port,
229 use "ovs-vsctl set", e.g.:
231 ovs-vsctl set port tap0 tag=9
233 Q: How do I configure a port as a SPAN port, that is, enable mirroring
234 of all traffic to that port?
236 A: The following commands configure br0 with eth0 and tap0 as trunk
237 ports. All traffic coming in or going out on eth0 or tap0 is also
238 mirrored to tap1; any traffic arriving on tap1 is dropped:
241 ovs-vsctl add-port br0 eth0
242 ovs-vsctl add-port br0 tap0
243 ovs-vsctl add-port br0 tap1 \
244 -- --id=@p get port tap1 \
245 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
246 -- set bridge br0 mirrors=@m
248 To later disable mirroring, run:
250 ovs-vsctl clear bridge br0 mirrors
252 Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable
253 mirroring of all traffic to that VLAN?
255 A: The following commands configure br0 with eth0 as a trunk port and
256 tap0 as an access port for VLAN 10. All traffic coming in or going
257 out on tap0, as well as traffic coming in or going out on eth0 in
258 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
259 VLAN 10, in cases where one is present, is dropped as part of
263 ovs-vsctl add-port br0 eth0
264 ovs-vsctl add-port br0 tap0 tag=10
266 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
268 -- set bridge br0 mirrors=@m
270 To later disable mirroring, run:
272 ovs-vsctl clear bridge br0 mirrors
274 Mirroring to a VLAN can disrupt a network that contains unmanaged
275 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
276 GRE tunnel has fewer caveats than mirroring to a VLAN and should
277 generally be preferred.
279 Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
281 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
282 of the specified output-vlan. This loss of information may make
283 the mirrored traffic too hard to interpret.
285 To mirror multiple VLANs, use the commands above, but specify a
286 comma-separated list of VLANs as the value for select-vlan. To
287 mirror every VLAN, use the commands above, but omit select-vlan and
290 When a packet arrives on a VLAN that is used as a mirror output
291 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
292 floods the packet across all the ports for which the mirror output
293 VLAN is configured. (If an OpenFlow controller is in use, then it
294 can override this behavior through the flow table.) If OVS is used
295 as an intermediate switch, rather than an edge switch, this ensures
296 that the RSPAN traffic is distributed through the network.
298 Mirroring to a VLAN can disrupt a network that contains unmanaged
299 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
300 GRE tunnel has fewer caveats than mirroring to a VLAN and should
301 generally be preferred.
303 Q: How do I configure mirroring of all traffic to a GRE tunnel?
305 A: The following commands configure br0 with eth0 and tap0 as trunk
306 ports. All traffic coming in or going out on eth0 or tap0 is also
307 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
308 traffic arriving on gre0 is dropped:
311 ovs-vsctl add-port br0 eth0
312 ovs-vsctl add-port br0 tap0
313 ovs-vsctl add-port br0 gre0 \
314 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
315 -- --id=@p get port gre0 \
316 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
317 -- set bridge br0 mirrors=@m
319 To later disable mirroring and destroy the GRE tunnel:
321 ovs-vsctl clear bridge br0 mirrors
322 ovs-vcstl del-port br0 gre0
324 Q: Does Open vSwitch support ERSPAN?
326 A: No. ERSPAN is an undocumented proprietary protocol. As an
327 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
330 Q: How do I connect two bridges?
332 A: First, why do you want to do this? Two connected bridges are not
333 much different from a single bridge, so you might as well just have
334 a single bridge with all your ports on it.
336 If you still want to connect two bridges, you can use a pair of
337 patch ports. The following example creates bridges br0 and br1,
338 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
339 and br1 with a pair of patch ports.
342 ovs-vsctl add-port br0 eth0
343 ovs-vsctl add-port br0 tap0
345 ovs-vsctl add-port br1 tap1
347 -- add-port br0 patch0 \
348 -- set interface patch0 type=patch options:peer=patch1 \
349 -- add-port br1 patch1 \
350 -- set interface patch1 type=patch options:peer=patch0
352 Bridges connected with patch ports are much like a single bridge.
353 For instance, if the example above also added eth1 to br1, and both
354 eth0 and eth1 happened to be connected to the same next-hop switch,
355 then you could loop your network just as you would if you added
356 eth0 and eth1 to the same bridge (see the "Configuration Problems"
357 section below for more information).
359 If you are using Open vSwitch 1.9 or an earlier version, then you
360 need to be using the kernel module bundled with Open vSwitch rather
361 than the one that is integrated into Linux 3.3 and later, because
362 Open vSwitch 1.9 and earlier versions need kernel support for patch
363 ports. This also means that in Open vSwitch 1.9 and earlier, patch
364 ports will not work with the userspace datapath, only with the
367 Q: Why are there so many different ways to dump flows?
369 A: Open vSwitch uses different kinds of flows for different purposes:
371 - OpenFlow flows are the most important kind of flow. OpenFlow
372 controllers use these flows to define a switch's policy.
373 OpenFlow flows support wildcards, priorities, and multiple
376 When in-band control is in use, Open vSwitch sets up a few
377 "hidden" flows, with priority higher than a controller or the
378 user can configure, that are not visible via OpenFlow. (See
379 the "Controller" section of the FAQ for more information
382 - The Open vSwitch software switch implementation uses a second
383 kind of flow internally. These flows, called "exact-match"
384 or "datapath" or "kernel" flows, do not support wildcards or
385 priorities and comprise only a single table, which makes them
386 suitable for caching. OpenFlow flows and exact-match flows
387 also support different actions and number ports differently.
389 Exact-match flows are an implementation detail that is
390 subject to change in future versions of Open vSwitch. Even
391 with the current version of Open vSwitch, hardware switch
392 implementations do not necessarily use exact-match flows.
394 Each of the commands for dumping flows has a different purpose:
396 - "ovs-ofctl dump-flows <br>" dumps OpenFlow flows, excluding
397 hidden flows. This is the most commonly useful form of flow
398 dump. (Unlike the other commands, this should work with any
399 OpenFlow switch, not just Open vSwitch.)
401 - "ovs-appctl bridge/dump-flows <br>" dumps OpenFlow flows,
402 including hidden flows. This is occasionally useful for
403 troubleshooting suspected issues with in-band control.
405 - "ovs-dpctl dump-flows [dp]" dumps the exact-match flow table
406 entries for a Linux kernel-based datapath. In Open vSwitch
407 1.10 and later, ovs-vswitchd merges multiple switches into a
408 single datapath, so it will show all the flows on all your
409 kernel-based switches. This command can occasionally be
410 useful for debugging.
412 - "ovs-appctl dpif/dump-flows <br>", new in Open vSwitch 1.10,
413 dumps exact-match flows for only the specified bridge,
414 regardless of the type.
417 Configuration Problems
418 ----------------------
420 Q: I created a bridge and added my Ethernet port to it, using commands
424 ovs-vsctl add-port br0 eth0
426 and as soon as I ran the "add-port" command I lost all connectivity
429 A: A physical Ethernet device that is part of an Open vSwitch bridge
430 should not have an IP address. If one does, then that IP address
431 will not be fully functional.
433 You can restore functionality by moving the IP address to an Open
434 vSwitch "internal" device, such as the network device named after
435 the bridge itself. For example, assuming that eth0's IP address is
436 192.168.128.5, you could run the commands below to fix up the
439 ifconfig eth0 0.0.0.0
440 ifconfig br0 192.168.128.5
442 (If your only connection to the machine running OVS is through the
443 IP address in question, then you would want to run all of these
444 commands on a single command line, or put them into a script.) If
445 there were any additional routes assigned to eth0, then you would
446 also want to use commands to adjust these routes to go through br0.
448 If you use DHCP to obtain an IP address, then you should kill the
449 DHCP client that was listening on the physical Ethernet interface
450 (e.g. eth0) and start one listening on the internal interface
451 (e.g. br0). You might still need to manually clear the IP address
452 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
454 There is no compelling reason why Open vSwitch must work this way.
455 However, this is the way that the Linux kernel bridge module has
456 always worked, so it's a model that those accustomed to Linux
457 bridging are already used to. Also, the model that most people
458 expect is not implementable without kernel changes on all the
459 versions of Linux that Open vSwitch supports.
461 By the way, this issue is not specific to physical Ethernet
462 devices. It applies to all network devices except Open vswitch
465 Q: I created a bridge and added a couple of Ethernet ports to it,
466 using commands like these:
469 ovs-vsctl add-port br0 eth0
470 ovs-vsctl add-port br0 eth1
472 and now my network seems to have melted: connectivity is unreliable
473 (even connectivity that doesn't go through Open vSwitch), all the
474 LEDs on my physical switches are blinking, wireshark shows
475 duplicated packets, and CPU usage is very high.
477 A: More than likely, you've looped your network. Probably, eth0 and
478 eth1 are connected to the same physical Ethernet switch. This
479 yields a scenario where OVS receives a broadcast packet on eth0 and
480 sends it out on eth1, then the physical switch connected to eth1
481 sends the packet back on eth0, and so on forever. More complicated
482 scenarios, involving a loop through multiple switches, are possible
485 The solution depends on what you are trying to do:
487 - If you added eth0 and eth1 to get higher bandwidth or higher
488 reliability between OVS and your physical Ethernet switch,
489 use a bond. The following commands create br0 and then add
490 eth0 and eth1 as a bond:
493 ovs-vsctl add-bond br0 bond0 eth0 eth1
495 Bonds have tons of configuration options. Please read the
496 documentation on the Port table in ovs-vswitchd.conf.db(5)
499 - Perhaps you don't actually need eth0 and eth1 to be on the
500 same bridge. For example, if you simply want to be able to
501 connect each of them to virtual machines, then you can put
502 each of them on a bridge of its own:
505 ovs-vsctl add-port br0 eth0
508 ovs-vsctl add-port br1 eth1
510 and then connect VMs to br0 and br1. (A potential
511 disadvantage is that traffic cannot directly pass between br0
512 and br1. Instead, it will go out eth0 and come back in eth1,
515 - If you have a redundant or complex network topology and you
516 want to prevent loops, turn on spanning tree protocol (STP).
517 The following commands create br0, enable STP, and add eth0
518 and eth1 to the bridge. The order is important because you
519 don't want have to have a loop in your network even
523 ovs-vsctl set bridge br0 stp_enable=true
524 ovs-vsctl add-port br0 eth0
525 ovs-vsctl add-port br0 eth1
527 The Open vSwitch implementation of STP is not well tested.
528 Please report any bugs you observe, but if you'd rather avoid
529 acting as a beta tester then another option might be your
532 Q: I can't seem to use Open vSwitch in a wireless network.
534 A: Wireless base stations generally only allow packets with the source
535 MAC address of NIC that completed the initial handshake.
536 Therefore, without MAC rewriting, only a single device can
537 communicate over a single wireless link.
539 This isn't specific to Open vSwitch, it's enforced by the access
540 point, so the same problems will show up with the Linux bridge or
541 any other way to do bridging.
543 Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
545 A: PPP most commonly carries IP packets, but Open vSwitch works only
546 with Ethernet frames. The correct way to interface PPP to an
547 Ethernet network is usually to use routing instead of switching.
549 Q: Is there any documentation on the database tables and fields?
551 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
553 Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
554 I only see a datapath called "ovs-system". How can I see datapath
555 information about a particular bridge?
557 A: In version 1.9.0, OVS switched to using a single datapath that is
558 shared by all bridges of that type. The "ovs-appctl dpif/*"
559 commands provide similar functionality that is scoped by the bridge.
562 Quality of Service (QoS)
563 ------------------------
565 Q: How do I configure Quality of Service (QoS)?
567 A: Suppose that you want to set up bridge br0 connected to physical
568 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
569 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
570 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
571 could configure the bridge this way:
575 add-port br0 eth0 -- \
576 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
577 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
578 set port eth0 qos=@newqos -- \
579 --id=@newqos create qos type=linux-htb \
580 other-config:max-rate=1000000000 \
581 queues:123=@vif10queue \
582 queues:234=@vif20queue -- \
583 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
584 --id=@vif20queue create queue other-config:max-rate=20000000
586 At this point, bridge br0 is configured with the ports and eth0 is
587 configured with the queues that you need for QoS, but nothing is
588 actually directing packets from vif1.0 or vif2.0 to the queues that
589 we have set up for them. That means that all of the packets to
590 eth0 are going to the "default queue", which is not what we want.
592 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
593 queues reserved for them:
595 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
596 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
598 Each of the above flows matches on the input port, sets up the
599 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
600 executes the "normal" action, which performs the same switching
601 that Open vSwitch would have done without any OpenFlow flows being
602 present. (We know that vif1.0 and vif2.0 have OpenFlow port
603 numbers 5 and 6, respectively, because we set their ofport_request
604 columns above. If we had not done that, then we would have needed
605 to find out their port numbers before setting up these flows.)
607 Now traffic going from vif1.0 or vif2.0 to eth0 should be
610 By the way, if you delete the bridge created by the above commands,
615 then that will leave one unreferenced QoS record and two
616 unreferenced Queue records in the Open vSwich database. One way to
617 clear them out, assuming you don't have other QoS or Queue records
618 that you want to keep, is:
620 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
622 If you do want to keep some QoS or Queue records, or the Open
623 vSwitch you are using is older than version 1.8 (which added the
624 --all option), then you will have to destroy QoS and Queue records
627 Q: I configured Quality of Service (QoS) in my OpenFlow network by
628 adding records to the QoS and Queue table, but the results aren't
631 A: Did you install OpenFlow flows that use your queues? This is the
632 primary way to tell Open vSwitch which queues you want to use. If
633 you don't do this, then the default queue will be used, which will
634 probably not have the effect you want.
636 Refer to the previous question for an example.
638 Q: I configured QoS, correctly, but my measurements show that it isn't
639 working as well as I expect.
641 A: With the Linux kernel, the Open vSwitch implementation of QoS has
644 - Open vSwitch configures a subset of Linux kernel QoS
645 features, according to what is in OVSDB. It is possible that
646 this code has bugs. If you believe that this is so, then you
647 can configure the Linux traffic control (QoS) stack directly
648 with the "tc" program. If you get better results that way,
649 you can send a detailed bug report to bugs@openvswitch.org.
651 It is certain that Open vSwitch cannot configure every Linux
652 kernel QoS feature. If you need some feature that OVS cannot
653 configure, then you can also use "tc" directly (or add that
656 - The Open vSwitch implementation of OpenFlow allows flows to
657 be directed to particular queues. This is pretty simple and
658 unlikely to have serious bugs at this point.
660 However, most problems with QoS on Linux are not bugs in Open
661 vSwitch at all. They tend to be either configuration errors
662 (please see the earlier questions in this section) or issues with
663 the traffic control (QoS) stack in Linux. The Open vSwitch
664 developers are not experts on Linux traffic control. We suggest
665 that, if you believe you are encountering a problem with Linux
666 traffic control, that you consult the tc manpages (e.g. tc(8),
667 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
668 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
676 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
677 partition a single switch into multiple switches. Suppose, for
678 example, that you have two groups of machines, group A and group B.
679 You want the machines in group A to be able to talk to each other,
680 and you want the machine in group B to be able to talk to each
681 other, but you don't want the machines in group A to be able to
682 talk to the machines in group B. You can do this with two
683 switches, by plugging the machines in group A into one switch and
684 the machines in group B into the other switch.
686 If you only have one switch, then you can use VLANs to do the same
687 thing, by configuring the ports for machines in group A as VLAN
688 "access ports" for one VLAN and the ports for group B as "access
689 ports" for a different VLAN. The switch will only forward packets
690 between ports that are assigned to the same VLAN, so this
691 effectively subdivides your single switch into two independent
692 switches, one for each group of machines.
694 So far we haven't said anything about VLAN headers. With access
695 ports, like we've described so far, no VLAN header is present in
696 the Ethernet frame. This means that the machines (or switches)
697 connected to access ports need not be aware that VLANs are
698 involved, just like in the case where we use two different physical
701 Now suppose that you have a whole bunch of switches in your
702 network, instead of just one, and that some machines in group A are
703 connected directly to both switches 1 and 2. To allow these
704 machines to talk to each other, you could add an access port for
705 group A's VLAN to switch 1 and another to switch 2, and then
706 connect an Ethernet cable between those ports. That works fine,
707 but it doesn't scale well as the number of switches and the number
708 of VLANs increases, because you use up a lot of valuable switch
709 ports just connecting together your VLANs.
711 This is where VLAN headers come in. Instead of using one cable and
712 two ports per VLAN to connect a pair of switches, we configure a
713 port on each switch as a VLAN "trunk port". Packets sent and
714 received on a trunk port carry a VLAN header that says what VLAN
715 the packet belongs to, so that only two ports total are required to
716 connect the switches, regardless of the number of VLANs in use.
717 Normally, only switches (either physical or virtual) are connected
718 to a trunk port, not individual hosts, because individual hosts
719 don't expect to see a VLAN header in the traffic that they receive.
721 None of the above discussion says anything about particular VLAN
722 numbers. This is because VLAN numbers are completely arbitrary.
723 One must only ensure that a given VLAN is numbered consistently
724 throughout a network and that different VLANs are given different
725 numbers. (That said, VLAN 0 is usually synonymous with a packet
726 that has no VLAN header, and VLAN 4095 is reserved.)
730 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
731 bugs. If you are having problems with VLANs that you suspect to be
732 driver related, then you have several options:
734 - Upgrade to Linux 3.3 or later.
736 - Build and install a fixed version of the particular driver
737 that is causing trouble, if one is available.
739 - Use a NIC whose driver does not have VLAN problems.
741 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
742 that works around bugs in kernel drivers. To enable VLAN
743 splinters on interface eth0, use the command:
745 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
747 For VLAN splinters to be effective, Open vSwitch must know
748 which VLANs are in use. See the "VLAN splinters" section in
749 the Interface table in ovs-vswitchd.conf.db(5) for details on
750 how Open vSwitch infers in-use VLANs.
752 VLAN splinters increase memory use and reduce performance, so
753 use them only if needed.
755 - Apply the "vlan workaround" patch from the XenServer kernel
756 patch queue, build Open vSwitch against this patched kernel,
757 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
758 workaround for each interface whose driver is buggy.
760 (This is a nontrivial exercise, so this option is included
761 only for completeness.)
763 It is not always easy to tell whether a Linux kernel driver has
764 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
765 can help you test. See their manpages for details. Of the two
766 utilities, ovs-test(8) is newer and more thorough, but
767 ovs-vlan-test(8) may be easier to use.
769 Q: VLANs still don't work. I've tested the driver so I know that it's OK.
771 A: Do you have VLANs enabled on the physical switch that OVS is
772 attached to? Make sure that the port is configured to trunk the
773 VLAN or VLANs that you are using with OVS.
775 Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
776 and to its destination host, but OVS seems to drop incoming return
779 A: It's possible that you have the VLAN configured on your physical
780 switch as the "native" VLAN. In this mode, the switch treats
781 incoming packets either tagged with the native VLAN or untagged as
782 part of the native VLAN. It may also send outgoing packets in the
783 native VLAN without a VLAN tag.
785 If this is the case, you have two choices:
787 - Change the physical switch port configuration to tag packets
788 it forwards to OVS with the native VLAN instead of forwarding
791 - Change the OVS configuration for the physical port to a
792 native VLAN mode. For example, the following sets up a
793 bridge with port eth0 in "native-tagged" mode in VLAN 9:
796 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
798 In this situation, "native-untagged" mode will probably work
799 equally well. Refer to the documentation for the Port table
800 in ovs-vswitchd.conf.db(5) for more information.
802 Q: I added a pair of VMs on different VLANs, like this:
805 ovs-vsctl add-port br0 eth0
806 ovs-vsctl add-port br0 tap0 tag=9
807 ovs-vsctl add-port br0 tap1 tag=10
809 but the VMs can't access each other, the external network, or the
812 A: It is to be expected that the VMs can't access each other. VLANs
813 are a means to partition a network. When you configured tap0 and
814 tap1 as access ports for different VLANs, you indicated that they
815 should be isolated from each other.
817 As for the external network and the Internet, it seems likely that
818 the machines you are trying to access are not on VLAN 9 (or 10) and
819 that the Internet is not available on VLAN 9 (or 10).
821 Q: I added a pair of VMs on the same VLAN, like this:
824 ovs-vsctl add-port br0 eth0
825 ovs-vsctl add-port br0 tap0 tag=9
826 ovs-vsctl add-port br0 tap1 tag=9
828 The VMs can access each other, but not the external network or the
831 A: It seems likely that the machines you are trying to access in the
832 external network are not on VLAN 9 and that the Internet is not
833 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
834 trunk VLAN on the upstream switch port to which eth0 is connected.
836 Q: Can I configure an IP address on a VLAN?
838 A: Yes. Use an "internal port" configured as an access port. For
839 example, the following configures IP address 192.168.0.7 on VLAN 9.
840 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
841 they have an 802.1Q header with VLAN 9. Conversely, traffic
842 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
846 ovs-vsctl add-port br0 eth0
847 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
848 ifconfig vlan9 192.168.0.7
850 Q: My OpenFlow controller doesn't see the VLANs that I expect.
852 A: The configuration for VLANs in the Open vSwitch database (e.g. via
853 ovs-vsctl) only affects traffic that goes through Open vSwitch's
854 implementation of the OpenFlow "normal switching" action. By
855 default, when Open vSwitch isn't connected to a controller and
856 nothing has been manually configured in the flow table, all traffic
857 goes through the "normal switching" action. But, if you set up
858 OpenFlow flows on your own, through a controller or using ovs-ofctl
859 or through other means, then you have to implement VLAN handling
862 You can use "normal switching" as a component of your OpenFlow
863 actions, e.g. by putting "normal" into the lists of actions on
864 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
865 controller. In situations where this is not suitable, you can
866 implement VLAN handling yourself, e.g.:
868 - If a packet comes in on an access port, and the flow table
869 needs to send it out on a trunk port, then the flow can add
870 the appropriate VLAN tag with the "mod_vlan_vid" action.
872 - If a packet comes in on a trunk port, and the flow table
873 needs to send it out on an access port, then the flow can
874 strip the VLAN tag with the "strip_vlan" action.
876 Q: I configured ports on a bridge as access ports with different VLAN
880 ovs-vsctl set-controller br0 tcp:192.168.0.10:6633
881 ovs-vsctl add-port br0 eth0
882 ovs-vsctl add-port br0 tap0 tag=9
883 ovs-vsctl add-port br0 tap1 tag=10
885 but the VMs running behind tap0 and tap1 can still communicate,
886 that is, they are not isolated from each other even though they are
889 A: Do you have a controller configured on br0 (as the commands above
890 do)? If so, then this is a variant on the previous question, "My
891 OpenFlow controller doesn't see the VLANs that I expect," and you
892 can refer to the answer there for more information.
900 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
901 to solve the scaling challenges of VLAN networks in a multi-tenant
902 environment. VXLAN is an overlay network which transports an L2 network
903 over an existing L3 network. For more information on VXLAN, please see
904 the IETF draft available here:
906 http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-03
908 Q: How much of the VXLAN protocol does Open vSwitch currently support?
910 A: Open vSwitch currently supports the framing format for packets on the
911 wire. There is currently no support for the multicast aspects of VXLAN.
912 To get around the lack of multicast support, it is possible to
913 pre-provision MAC to IP address mappings either manually or from a
916 Q: What destination UDP port does the VXLAN implementation in Open vSwitch
919 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
920 is 4789. However, it is possible to configure the destination UDP port
921 manually on a per-VXLAN tunnel basis. An example of this configuration is
925 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
926 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
927 options:dst_port=8472
930 Using OpenFlow (Manually or Via Controller)
931 -------------------------------------------
933 Q: What versions of OpenFlow does Open vSwitch support?
935 A: Open vSwitch 1.9 and earlier support only OpenFlow 1.0 (plus
936 extensions that bring in many of the features from later versions
939 Open vSwitch versions 1.10 and later will have experimental support
940 for OpenFlow 1.2 and 1.3. On these versions of Open vSwitch, the
941 following command enables OpenFlow 1.0, 1.2, and 1.3 on bridge br0:
943 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow12,OpenFlow13
945 Support for OpenFlow 1.1 is incomplete enough that it cannot yet be
946 enabled, even experimentally.
948 Support for OpenFlow 1.2 and 1.3 is still incomplete. Work to be
949 done is tracked in OPENFLOW-1.1+ in the Open vSwitch source tree
950 (also via http://openvswitch.org/development/openflow-1-x-plan/).
951 When support for a given OpenFlow version is solidly implemented,
952 Open vSwitch will enable that version by default.
954 Q: I'm getting "error type 45250 code 0". What's that?
956 A: This is a Open vSwitch extension to OpenFlow error codes. Open
957 vSwitch uses this extension when it must report an error to an
958 OpenFlow controller but no standard OpenFlow error code is
961 Open vSwitch logs the errors that it sends to controllers, so the
962 easiest thing to do is probably to look at the ovs-vswitchd log to
963 find out what the error was.
965 If you want to dissect the extended error message yourself, the
966 format is documented in include/openflow/nicira-ext.h in the Open
967 vSwitch source distribution. The extended error codes are
968 documented in lib/ofp-errors.h.
970 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
971 doesn't actually appear through the OpenFlow connection, even
972 though I know that it's going through.
973 Q2: Some of the OpenFlow flows that my controller sets up don't seem
974 to apply to certain traffic, especially traffic between OVS and
975 the controller itself.
977 A: By default, Open vSwitch assumes that OpenFlow controllers are
978 connected "in-band", that is, that the controllers are actually
979 part of the network that is being controlled. In in-band mode,
980 Open vSwitch sets up special "hidden" flows to make sure that
981 traffic can make it back and forth between OVS and the controllers.
982 These hidden flows are higher priority than any flows that can be
983 set up through OpenFlow, and they are not visible through normal
984 OpenFlow flow table dumps.
986 Usually, the hidden flows are desirable and helpful, but
987 occasionally they can cause unexpected behavior. You can view the
988 full OpenFlow flow table, including hidden flows, on bridge br0
991 ovs-appctl bridge/dump-flows br0
993 to help you debug. The hidden flows are those with priorities
994 greater than 65535 (the maximum priority that can be set with
997 The DESIGN file at the top level of the Open vSwitch source
998 distribution describes the in-band model in detail.
1000 If your controllers are not actually in-band (e.g. they are on
1001 localhost via 127.0.0.1, or on a separate network), then you should
1002 configure your controllers in "out-of-band" mode. If you have one
1003 controller on bridge br0, then you can configure out-of-band mode
1006 ovs-vsctl set controller br0 connection-mode=out-of-band
1008 Q: I configured all my controllers for out-of-band control mode but
1009 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1011 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1012 set-manager"). By default, Open vSwitch assumes that managers need
1013 in-band rules set up on every bridge. You can disable these rules
1016 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1018 This actually disables in-band control entirely for the bridge, as
1019 if all the bridge's controllers were configured for out-of-band
1022 Q: My OpenFlow controller doesn't see the VLANs that I expect.
1024 A: See answer under "VLANs", above.
1026 Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1027 but I got a funny message like this:
1029 ofp_util|INFO|normalization changed ofp_match, details:
1030 ofp_util|INFO| pre: nw_dst=192.168.0.1
1033 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1034 match had disappeared, so that the flow ends up matching every
1037 A: The term "normalization" in the log message means that a flow
1038 cannot match on an L3 field without saying what L3 protocol is in
1039 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1040 so the L3 field match was dropped.
1042 In this case, the L3 protocol could be IP or ARP. A correct
1043 command for each possibility is, respectively:
1045 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1049 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1051 Similarly, a flow cannot match on an L4 field without saying what
1052 L4 protocol is in use. For example, the flow match "tp_src=1234"
1053 is, by itself, meaningless and will be ignored. Instead, to match
1054 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1055 source port 1234, write "udp,tp_src=1234".
1057 Q: How can I figure out the OpenFlow port number for a given port?
1059 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1060 respond with an OFPT_FEATURES_REPLY that, among other information,
1061 includes a mapping between OpenFlow port names and numbers. From a
1062 command prompt, "ovs-ofctl show br0" makes such a request and
1063 prints the response for switch br0.
1065 The Interface table in the Open vSwitch database also maps OpenFlow
1066 port names to numbers. To print the OpenFlow port number
1067 associated with interface eth0, run:
1069 ovs-vsctl get Interface eth0 ofport
1071 You can print the entire mapping with:
1073 ovs-vsctl -- --columns=name,ofport list Interface
1075 but the output mixes together interfaces from all bridges in the
1076 database, so it may be confusing if more than one bridge exists.
1078 In the Open vSwitch database, ofport value -1 means that the
1079 interface could not be created due to an error. (The Open vSwitch
1080 log should indicate the reason.) ofport value [] (the empty set)
1081 means that the interface hasn't been created yet. The latter is
1082 normally an intermittent condition (unless ovs-vswitchd is not
1085 Q: I added some flows with my controller or with ovs-ofctl, but when I
1086 run "ovs-dpctl dump-flows" I don't see them.
1088 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1089 won't display the information that you want. You want to use
1090 "ovs-ofctl dump-flows" instead.
1092 Q: It looks like each of the interfaces in my bonded port shows up
1093 as an individual OpenFlow port. Is that right?
1095 A: Yes, Open vSwitch makes individual bond interfaces visible as
1096 OpenFlow ports, rather than the bond as a whole. The interfaces
1097 are treated together as a bond for only a few purposes:
1099 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1100 controller is not configured, this happens implicitly to
1103 - Mirrors configured for output to a bonded port.
1105 It would make a lot of sense for Open vSwitch to present a bond as
1106 a single OpenFlow port. If you want to contribute an
1107 implementation of such a feature, please bring it up on the Open
1108 vSwitch development mailing list at dev@openvswitch.org.
1110 Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1111 multiple hosts, and other components. The behavior isn't what I
1114 A: To debug network behavior problems, trace the path of a packet,
1115 hop-by-hop, from its origin in one host to a remote host. If
1116 that's correct, then trace the path of the response packet back to
1119 Usually a simple ICMP echo request and reply ("ping") packet is
1120 good enough. Start by initiating an ongoing "ping" from the origin
1121 host to a remote host. If you are tracking down a connectivity
1122 problem, the "ping" will not display any successful output, but
1123 packets are still being sent. (In this case the packets being sent
1124 are likely ARP rather than ICMP.)
1126 Tools available for tracing include the following:
1128 - "tcpdump" and "wireshark" for observing hops across network
1129 devices, such as Open vSwitch internal devices and physical
1132 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1133 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1134 These tools allow one to observe the actions being taken on
1135 packets in ongoing flows.
1137 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1138 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1139 documentation, and "Why are there so many different ways to
1140 dump flows?" above for some background.
1142 - "ovs-appctl ofproto/trace" to observe the logic behind how
1143 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1144 documentation. You can out more details about a given flow
1145 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1146 a flow from the output into an "ovs-appctl ofproto/trace"
1149 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1150 observe what goes on at these physical hops.
1152 Starting at the origin of a given packet, observe the packet at
1153 each hop in turn. For example, in one plausible scenario, you
1156 1. "tcpdump" the "eth" interface through which an ARP egresses
1157 a VM, from inside the VM.
1159 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1160 ingresses the host machine.
1162 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1163 the host interface through which the ARP egresses the
1164 physical machine. You may need to use "ovs-dpctl show" to
1165 interpret the port numbers. If the output seems surprising,
1166 you can use "ovs-appctl ofproto/trace" to observe details of
1167 how ovs-vswitchd determined the actions in the "ovs-dpctl
1170 4. "tcpdump" the "eth" interface through which the ARP egresses
1171 the physical machine.
1173 5. "tcpdump" the "eth" interface through which the ARP
1174 ingresses the physical machine, at the remote host that
1177 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1178 remote host that receives the ARP and observe the VM "vif"
1179 or "tap" interface to which the flow is directed. Again,
1180 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1182 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1185 8. "tcpdump" the "eth" interface through which the ARP
1186 ingresses a VM, from inside the VM.
1188 It is likely that during one of these steps you will figure out the
1189 problem. If not, then follow the ARP reply back to the origin, in
1192 Q: How do I make a flow drop packets?
1194 A: An empty set of actions causes a packet to be dropped. You can
1195 specify an empty set of actions with "actions=" on the ovs-ofctl
1196 command line. For example:
1198 ovs-ofctl add-flow br0 priority=65535,actions=
1200 would cause every packet entering switch br0 to be dropped.
1202 You can write "drop" explicitly if you like. The effect is the
1203 same. Thus, the following command also causes every packet
1204 entering switch br0 to be dropped:
1206 ovs-ofctl add-flow br0 priority=65535,actions=drop
1212 bugs@openvswitch.org
1213 http://openvswitch.org/