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 environments. A
13 vswitch forwards traffic between different VMs on the same physical host
14 and also forwards traffic between VMs and the physical network. Open
15 vSwitch supports standard management interfaces (e.g. sFlow, NetFlow,
16 RSPAN, CLI), and is open to programmatic extension and control using
17 OpenFlow and the OVSDB management protocol.
19 Open vSwitch as designed to be compatible with modern switching
20 chipsets. This means that it can be ported to existing high-fanout
21 switches allowing the same flexible control of the physical
22 infrastructure as the virtual infrastructure. It also means that
23 Open vSwitch will be able to take advantage of on-NIC switching
24 chipsets as their functionality matures.
26 Q: What virtualization platforms can use Open vSwitch?
28 A: Open vSwitch can currently run on any Linux-based virtualization
29 platform (kernel 2.6.18 and newer), including: KVM, VirtualBox, Xen,
30 Xen Cloud Platform, XenServer. As of Linux 3.3 it is part of the
31 mainline kernel. The bulk of the code is written in platform-
32 independent C and is easily ported to other environments. We welcome
33 inquires about integrating Open vSwitch with other virtualization
36 Q: How can I try Open vSwitch?
38 A: The Open vSwitch source code can be built on a Linux system. You can
39 build and experiment with Open vSwitch on any Linux machine.
40 Packages for various Linux distributions are available on many
41 platforms, including: Debian, Ubuntu, Fedora.
43 You may also download and run a virtualization platform that already
44 has Open vSwitch integrated. For example, download a recent ISO for
45 XenServer or Xen Cloud Platform. Be aware that the version
46 integrated with a particular platform may not be the most recent Open
49 Q: Does Open vSwitch only work on Linux?
51 A: No, Open vSwitch has been ported to a number of different operating
52 systems and hardware platforms. Most of the development work occurs
53 on Linux, but the code should be portable to any POSIX system. We've
54 seen Open vSwitch ported to a number of different platforms,
55 including FreeBSD, Windows, and even non-POSIX embedded systems.
57 By definition, the Open vSwitch Linux kernel module only works on
58 Linux and will provide the highest performance. However, a userspace
59 datapath is available that should be very portable.
61 Q: What's involved with porting Open vSwitch to a new platform or
64 A: The PORTING document describes how one would go about porting Open
65 vSwitch to a new operating system or hardware platform.
67 Q: Why would I use Open vSwitch instead of the Linux bridge?
69 A: Open vSwitch is specially designed to make it easier to manage VM
70 network configuration and monitor state spread across many physical
71 hosts in dynamic virtualized environments. Please see WHY-OVS for a
72 more detailed description of how Open vSwitch relates to the Linux
75 Q: How is Open vSwitch related to distributed virtual switches like the
76 VMware vNetwork distributed switch or the Cisco Nexus 1000V?
78 A: Distributed vswitch applications (e.g., VMware vNetwork distributed
79 switch, Cisco Nexus 1000V) provide a centralized way to configure and
80 monitor the network state of VMs that are spread across many physical
81 hosts. Open vSwitch is not a distributed vswitch itself, rather it
82 runs on each physical host and supports remote management in a way
83 that makes it easier for developers of virtualization/cloud
84 management platforms to offer distributed vswitch capabilities.
86 To aid in distribution, Open vSwitch provides two open protocols that
87 are specially designed for remote management in virtualized network
88 environments: OpenFlow, which exposes flow-based forwarding state,
89 and the OVSDB management protocol, which exposes switch port state.
90 In addition to the switch implementation itself, Open vSwitch
91 includes tools (ovs-controller, ovs-ofctl, ovs-vsctl) that developers
92 can script and extend to provide distributed vswitch capabilities
93 that are closely integrated with their virtualization management
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 ------------ -------------
149 Open vSwitch userspace should also work with the Linux kernel module
150 built into Linux 3.3 and later.
152 Open vSwitch userspace is not sensitive to the Linux kernel version.
153 It should build against almost any kernel, certainly against 2.6.18
156 Q: Should userspace or kernel be upgraded first to minimize downtime?
158 In general, the Open vSwitch userspace should be used with the
159 kernel version included in the same release or with the version
160 from upstream Linux. However, when upgrading between two releases
161 of Open vSwitch it is best to migrate userspace first to reduce
162 the possbility of incompatibilities.
164 Q: What features are not available in the Open vSwitch kernel datapath
165 that ships as part of the upstream Linux kernel?
167 A: The kernel module in upstream Linux 3.3 and later does not include
168 tunnel virtual ports, that is, interfaces with type "gre",
169 "ipsec_gre", "gre64", "ipsec_gre64", "vxlan", or "lisp". It is
170 possible to create tunnels in Linux and attach them to Open vSwitch
171 as system devices. However, they cannot be dynamically created
172 through the OVSDB protocol or set the tunnel ids as a flow action.
174 Work is in progress in adding tunnel virtual ports to the upstream
175 Linux version of the Open vSwitch kernel module. For now, if you
176 need these features, use the kernel module from the Open vSwitch
177 distribution instead of the upstream Linux kernel module.
179 The upstream kernel module does not include patch ports, but this
180 only matters for Open vSwitch 1.9 and earlier, because Open vSwitch
181 1.10 and later implement patch ports without using this kernel
184 Q: What features are not available when using the userspace datapath?
186 A: Tunnel virtual ports are not supported, as described in the
187 previous answer. It is also not possible to use queue-related
188 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
189 may not be transmitted.
195 Q: I thought Open vSwitch was a virtual Ethernet switch, but the
196 documentation keeps talking about bridges. What's a bridge?
198 A: In networking, the terms "bridge" and "switch" are synonyms. Open
199 vSwitch implements an Ethernet switch, which means that it is also
204 A: See the "VLAN" section below.
210 Q: How do I configure a port as an access port?
212 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
213 the following commands configure br0 with eth0 as a trunk port (the
214 default) and tap0 as an access port for VLAN 9:
217 ovs-vsctl add-port br0 eth0
218 ovs-vsctl add-port br0 tap0 tag=9
220 If you want to configure an already added port as an access port,
221 use "ovs-vsctl set", e.g.:
223 ovs-vsctl set port tap0 tag=9
225 Q: How do I configure a port as a SPAN port, that is, enable mirroring
226 of all traffic to that port?
228 A: The following commands configure br0 with eth0 and tap0 as trunk
229 ports. All traffic coming in or going out on eth0 or tap0 is also
230 mirrored to tap1; any traffic arriving on tap1 is dropped:
233 ovs-vsctl add-port br0 eth0
234 ovs-vsctl add-port br0 tap0
235 ovs-vsctl add-port br0 tap1 \
236 -- --id=@p get port tap1 \
237 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
238 -- set bridge br0 mirrors=@m
240 To later disable mirroring, run:
242 ovs-vsctl clear bridge br0 mirrors
244 Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable
245 mirroring of all traffic to that VLAN?
247 A: The following commands configure br0 with eth0 as a trunk port and
248 tap0 as an access port for VLAN 10. All traffic coming in or going
249 out on tap0, as well as traffic coming in or going out on eth0 in
250 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
251 VLAN 10, in cases where one is present, is dropped as part of
255 ovs-vsctl add-port br0 eth0
256 ovs-vsctl add-port br0 tap0 tag=10
258 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
260 -- set bridge br0 mirrors=@m
262 To later disable mirroring, run:
264 ovs-vsctl clear bridge br0 mirrors
266 Mirroring to a VLAN can disrupt a network that contains unmanaged
267 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
268 GRE tunnel has fewer caveats than mirroring to a VLAN and should
269 generally be preferred.
271 Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
273 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
274 of the specified output-vlan. This loss of information may make
275 the mirrored traffic too hard to interpret.
277 To mirror multiple VLANs, use the commands above, but specify a
278 comma-separated list of VLANs as the value for select-vlan. To
279 mirror every VLAN, use the commands above, but omit select-vlan and
282 When a packet arrives on a VLAN that is used as a mirror output
283 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
284 floods the packet across all the ports for which the mirror output
285 VLAN is configured. (If an OpenFlow controller is in use, then it
286 can override this behavior through the flow table.) If OVS is used
287 as an intermediate switch, rather than an edge switch, this ensures
288 that the RSPAN traffic is distributed through the network.
290 Mirroring to a VLAN can disrupt a network that contains unmanaged
291 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
292 GRE tunnel has fewer caveats than mirroring to a VLAN and should
293 generally be preferred.
295 Q: How do I configure mirroring of all traffic to a GRE tunnel?
297 A: The following commands configure br0 with eth0 and tap0 as trunk
298 ports. All traffic coming in or going out on eth0 or tap0 is also
299 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
300 traffic arriving on gre0 is dropped:
303 ovs-vsctl add-port br0 eth0
304 ovs-vsctl add-port br0 tap0
305 ovs-vsctl add-port br0 gre0 \
306 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
307 -- --id=@p get port gre0 \
308 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
309 -- set bridge br0 mirrors=@m
311 To later disable mirroring and destroy the GRE tunnel:
313 ovs-vsctl clear bridge br0 mirrors
314 ovs-vcstl del-port br0 gre0
316 Q: Does Open vSwitch support ERSPAN?
318 A: No. ERSPAN is an undocumented proprietary protocol. As an
319 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
322 Q: Why are there so many different ways to dump flows?
324 A: Open vSwitch uses different kinds of flows for different purposes:
326 - OpenFlow flows are the most important kind of flow. OpenFlow
327 controllers use these flows to define a switch's policy.
328 OpenFlow flows support wildcards, priorities, and multiple
331 When in-band control is in use, Open vSwitch sets up a few
332 "hidden" flows, with priority higher than a controller or the
333 user can configure, that are not visible via OpenFlow. (See
334 the "Controller" section of the FAQ for more information
337 - The Open vSwitch software switch implementation uses a second
338 kind of flow internally. These flows, called "exact-match"
339 or "datapath" or "kernel" flows, do not support wildcards or
340 priorities and comprise only a single table, which makes them
341 suitable for caching. OpenFlow flows and exact-match flows
342 also support different actions and number ports differently.
344 Exact-match flows are an implementation detail that is
345 subject to change in future versions of Open vSwitch. Even
346 with the current version of Open vSwitch, hardware switch
347 implementations do not necessarily use exact-match flows.
349 Each of the commands for dumping flows has a different purpose:
351 - "ovs-ofctl dump-flows <br>" dumps OpenFlow flows, excluding
352 hidden flows. This is the most commonly useful form of flow
353 dump. (Unlike the other commands, this should work with any
354 OpenFlow switch, not just Open vSwitch.)
356 - "ovs-appctl bridge/dump-flows <br>" dumps OpenFlow flows,
357 including hidden flows. This is occasionally useful for
358 troubleshooting suspected issues with in-band control.
360 - "ovs-dpctl dump-flows [dp]" dumps the exact-match flow table
361 entries for a Linux kernel-based datapath. In Open vSwitch
362 1.10 and later, ovs-vswitchd merges multiple switches into a
363 single datapath, so it will show all the flows on all your
364 kernel-based switches. This command can occasionally be
365 useful for debugging.
367 - "ovs-appctl dpif/dump-flows <br>", new in Open vSwitch 1.10,
368 dumps exact-match flows for only the specified bridge,
369 regardless of the type.
372 Configuration Problems
373 ----------------------
375 Q: I created a bridge and added my Ethernet port to it, using commands
379 ovs-vsctl add-port br0 eth0
381 and as soon as I ran the "add-port" command I lost all connectivity
384 A: A physical Ethernet device that is part of an Open vSwitch bridge
385 should not have an IP address. If one does, then that IP address
386 will not be fully functional.
388 You can restore functionality by moving the IP address to an Open
389 vSwitch "internal" device, such as the network device named after
390 the bridge itself. For example, assuming that eth0's IP address is
391 192.168.128.5, you could run the commands below to fix up the
394 ifconfig eth0 0.0.0.0
395 ifconfig br0 192.168.128.5
397 (If your only connection to the machine running OVS is through the
398 IP address in question, then you would want to run all of these
399 commands on a single command line, or put them into a script.) If
400 there were any additional routes assigned to eth0, then you would
401 also want to use commands to adjust these routes to go through br0.
403 If you use DHCP to obtain an IP address, then you should kill the
404 DHCP client that was listening on the physical Ethernet interface
405 (e.g. eth0) and start one listening on the internal interface
406 (e.g. br0). You might still need to manually clear the IP address
407 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
409 There is no compelling reason why Open vSwitch must work this way.
410 However, this is the way that the Linux kernel bridge module has
411 always worked, so it's a model that those accustomed to Linux
412 bridging are already used to. Also, the model that most people
413 expect is not implementable without kernel changes on all the
414 versions of Linux that Open vSwitch supports.
416 By the way, this issue is not specific to physical Ethernet
417 devices. It applies to all network devices except Open vswitch
420 Q: I created a bridge and added a couple of Ethernet ports to it,
421 using commands like these:
424 ovs-vsctl add-port br0 eth0
425 ovs-vsctl add-port br0 eth1
427 and now my network seems to have melted: connectivity is unreliable
428 (even connectivity that doesn't go through Open vSwitch), all the
429 LEDs on my physical switches are blinking, wireshark shows
430 duplicated packets, and CPU usage is very high.
432 A: More than likely, you've looped your network. Probably, eth0 and
433 eth1 are connected to the same physical Ethernet switch. This
434 yields a scenario where OVS receives a broadcast packet on eth0 and
435 sends it out on eth1, then the physical switch connected to eth1
436 sends the packet back on eth0, and so on forever. More complicated
437 scenarios, involving a loop through multiple switches, are possible
440 The solution depends on what you are trying to do:
442 - If you added eth0 and eth1 to get higher bandwidth or higher
443 reliability between OVS and your physical Ethernet switch,
444 use a bond. The following commands create br0 and then add
445 eth0 and eth1 as a bond:
448 ovs-vsctl add-bond br0 bond0 eth0 eth1
450 Bonds have tons of configuration options. Please read the
451 documentation on the Port table in ovs-vswitchd.conf.db(5)
454 - Perhaps you don't actually need eth0 and eth1 to be on the
455 same bridge. For example, if you simply want to be able to
456 connect each of them to virtual machines, then you can put
457 each of them on a bridge of its own:
460 ovs-vsctl add-port br0 eth0
463 ovs-vsctl add-port br1 eth1
465 and then connect VMs to br0 and br1. (A potential
466 disadvantage is that traffic cannot directly pass between br0
467 and br1. Instead, it will go out eth0 and come back in eth1,
470 - If you have a redundant or complex network topology and you
471 want to prevent loops, turn on spanning tree protocol (STP).
472 The following commands create br0, enable STP, and add eth0
473 and eth1 to the bridge. The order is important because you
474 don't want have to have a loop in your network even
478 ovs-vsctl set bridge br0 stp_enable=true
479 ovs-vsctl add-port br0 eth0
480 ovs-vsctl add-port br0 eth1
482 The Open vSwitch implementation of STP is not well tested.
483 Please report any bugs you observe, but if you'd rather avoid
484 acting as a beta tester then another option might be your
487 Q: I can't seem to use Open vSwitch in a wireless network.
489 A: Wireless base stations generally only allow packets with the source
490 MAC address of NIC that completed the initial handshake.
491 Therefore, without MAC rewriting, only a single device can
492 communicate over a single wireless link.
494 This isn't specific to Open vSwitch, it's enforced by the access
495 point, so the same problems will show up with the Linux bridge or
496 any other way to do bridging.
498 Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
500 A: PPP most commonly carries IP packets, but Open vSwitch works only
501 with Ethernet frames. The correct way to interface PPP to an
502 Ethernet network is usually to use routing instead of switching.
504 Q: Is there any documentation on the database tables and fields?
506 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
508 Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
509 I only see a datapath called "ovs-system". How can I see datapath
510 information about a particular bridge?
512 A: In version 1.9.0, OVS switched to using a single datapath that is
513 shared by all bridges of that type. The "ovs-appctl dpif/*"
514 commands provide similar functionality that is scoped by the bridge.
517 Quality of Service (QoS)
518 ------------------------
520 Q: How do I configure Quality of Service (QoS)?
522 A: Suppose that you want to set up bridge br0 connected to physical
523 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
524 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
525 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
526 could configure the bridge this way:
530 add-port br0 eth0 -- \
531 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
532 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
533 set port eth0 qos=@newqos -- \
534 --id=@newqos create qos type=linux-htb \
535 other-config:max-rate=1000000000 \
536 queues:123=@vif10queue \
537 queues:234=@vif20queue -- \
538 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
539 --id=@vif20queue create queue other-config:max-rate=20000000
541 At this point, bridge br0 is configured with the ports and eth0 is
542 configured with the queues that you need for QoS, but nothing is
543 actually directing packets from vif1.0 or vif2.0 to the queues that
544 we have set up for them. That means that all of the packets to
545 eth0 are going to the "default queue", which is not what we want.
547 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
548 queues reserved for them:
550 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
551 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
553 Each of the above flows matches on the input port, sets up the
554 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
555 executes the "normal" action, which performs the same switching
556 that Open vSwitch would have done without any OpenFlow flows being
557 present. (We know that vif1.0 and vif2.0 have OpenFlow port
558 numbers 5 and 6, respectively, because we set their ofport_request
559 columns above. If we had not done that, then we would have needed
560 to find out their port numbers before setting up these flows.)
562 Now traffic going from vif1.0 or vif2.0 to eth0 should be
565 By the way, if you delete the bridge created by the above commands,
570 then that will leave one unreferenced QoS record and two
571 unreferenced Queue records in the Open vSwich database. One way to
572 clear them out, assuming you don't have other QoS or Queue records
573 that you want to keep, is:
575 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
577 If you do want to keep some QoS or Queue records, or the Open
578 vSwitch you are using is older than version 1.8 (which added the
579 --all option), then you will have to destroy QoS and Queue records
582 Q: I configured Quality of Service (QoS) in my OpenFlow network by
583 adding records to the QoS and Queue table, but the results aren't
586 A: Did you install OpenFlow flows that use your queues? This is the
587 primary way to tell Open vSwitch which queues you want to use. If
588 you don't do this, then the default queue will be used, which will
589 probably not have the effect you want.
591 Refer to the previous question for an example.
593 Q: I configured QoS, correctly, but my measurements show that it isn't
594 working as well as I expect.
596 A: With the Linux kernel, the Open vSwitch implementation of QoS has
599 - Open vSwitch configures a subset of Linux kernel QoS
600 features, according to what is in OVSDB. It is possible that
601 this code has bugs. If you believe that this is so, then you
602 can configure the Linux traffic control (QoS) stack directly
603 with the "tc" program. If you get better results that way,
604 you can send a detailed bug report to bugs@openvswitch.org.
606 It is certain that Open vSwitch cannot configure every Linux
607 kernel QoS feature. If you need some feature that OVS cannot
608 configure, then you can also use "tc" directly (or add that
611 - The Open vSwitch implementation of OpenFlow allows flows to
612 be directed to particular queues. This is pretty simple and
613 unlikely to have serious bugs at this point.
615 However, most problems with QoS on Linux are not bugs in Open
616 vSwitch at all. They tend to be either configuration errors
617 (please see the earlier questions in this section) or issues with
618 the traffic control (QoS) stack in Linux. The Open vSwitch
619 developers are not experts on Linux traffic control. We suggest
620 that, if you believe you are encountering a problem with Linux
621 traffic control, that you consult the tc manpages (e.g. tc(8),
622 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
623 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
631 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
632 partition a single switch into multiple switches. Suppose, for
633 example, that you have two groups of machines, group A and group B.
634 You want the machines in group A to be able to talk to each other,
635 and you want the machine in group B to be able to talk to each
636 other, but you don't want the machines in group A to be able to
637 talk to the machines in group B. You can do this with two
638 switches, by plugging the machines in group A into one switch and
639 the machines in group B into the other switch.
641 If you only have one switch, then you can use VLANs to do the same
642 thing, by configuring the ports for machines in group A as VLAN
643 "access ports" for one VLAN and the ports for group B as "access
644 ports" for a different VLAN. The switch will only forward packets
645 between ports that are assigned to the same VLAN, so this
646 effectively subdivides your single switch into two independent
647 switches, one for each group of machines.
649 So far we haven't said anything about VLAN headers. With access
650 ports, like we've described so far, no VLAN header is present in
651 the Ethernet frame. This means that the machines (or switches)
652 connected to access ports need not be aware that VLANs are
653 involved, just like in the case where we use two different physical
656 Now suppose that you have a whole bunch of switches in your
657 network, instead of just one, and that some machines in group A are
658 connected directly to both switches 1 and 2. To allow these
659 machines to talk to each other, you could add an access port for
660 group A's VLAN to switch 1 and another to switch 2, and then
661 connect an Ethernet cable between those ports. That works fine,
662 but it doesn't scale well as the number of switches and the number
663 of VLANs increases, because you use up a lot of valuable switch
664 ports just connecting together your VLANs.
666 This is where VLAN headers come in. Instead of using one cable and
667 two ports per VLAN to connect a pair of switches, we configure a
668 port on each switch as a VLAN "trunk port". Packets sent and
669 received on a trunk port carry a VLAN header that says what VLAN
670 the packet belongs to, so that only two ports total are required to
671 connect the switches, regardless of the number of VLANs in use.
672 Normally, only switches (either physical or virtual) are connected
673 to a trunk port, not individual hosts, because individual hosts
674 don't expect to see a VLAN header in the traffic that they receive.
676 None of the above discussion says anything about particular VLAN
677 numbers. This is because VLAN numbers are completely arbitrary.
678 One must only ensure that a given VLAN is numbered consistently
679 throughout a network and that different VLANs are given different
680 numbers. (That said, VLAN 0 is usually synonymous with a packet
681 that has no VLAN header, and VLAN 4095 is reserved.)
685 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
686 bugs. If you are having problems with VLANs that you suspect to be
687 driver related, then you have several options:
689 - Upgrade to Linux 3.3 or later.
691 - Build and install a fixed version of the particular driver
692 that is causing trouble, if one is available.
694 - Use a NIC whose driver does not have VLAN problems.
696 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
697 that works around bugs in kernel drivers. To enable VLAN
698 splinters on interface eth0, use the command:
700 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
702 For VLAN splinters to be effective, Open vSwitch must know
703 which VLANs are in use. See the "VLAN splinters" section in
704 the Interface table in ovs-vswitchd.conf.db(5) for details on
705 how Open vSwitch infers in-use VLANs.
707 VLAN splinters increase memory use and reduce performance, so
708 use them only if needed.
710 - Apply the "vlan workaround" patch from the XenServer kernel
711 patch queue, build Open vSwitch against this patched kernel,
712 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
713 workaround for each interface whose driver is buggy.
715 (This is a nontrivial exercise, so this option is included
716 only for completeness.)
718 It is not always easy to tell whether a Linux kernel driver has
719 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
720 can help you test. See their manpages for details. Of the two
721 utilities, ovs-test(8) is newer and more thorough, but
722 ovs-vlan-test(8) may be easier to use.
724 Q: VLANs still don't work. I've tested the driver so I know that it's OK.
726 A: Do you have VLANs enabled on the physical switch that OVS is
727 attached to? Make sure that the port is configured to trunk the
728 VLAN or VLANs that you are using with OVS.
730 Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
731 and to its destination host, but OVS seems to drop incoming return
734 A: It's possible that you have the VLAN configured on your physical
735 switch as the "native" VLAN. In this mode, the switch treats
736 incoming packets either tagged with the native VLAN or untagged as
737 part of the native VLAN. It may also send outgoing packets in the
738 native VLAN without a VLAN tag.
740 If this is the case, you have two choices:
742 - Change the physical switch port configuration to tag packets
743 it forwards to OVS with the native VLAN instead of forwarding
746 - Change the OVS configuration for the physical port to a
747 native VLAN mode. For example, the following sets up a
748 bridge with port eth0 in "native-tagged" mode in VLAN 9:
751 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
753 In this situation, "native-untagged" mode will probably work
754 equally well. Refer to the documentation for the Port table
755 in ovs-vswitchd.conf.db(5) for more information.
757 Q: I added a pair of VMs on different VLANs, like this:
760 ovs-vsctl add-port br0 eth0
761 ovs-vsctl add-port br0 tap0 tag=9
762 ovs-vsctl add-port br0 tap1 tag=10
764 but the VMs can't access each other, the external network, or the
767 A: It is to be expected that the VMs can't access each other. VLANs
768 are a means to partition a network. When you configured tap0 and
769 tap1 as access ports for different VLANs, you indicated that they
770 should be isolated from each other.
772 As for the external network and the Internet, it seems likely that
773 the machines you are trying to access are not on VLAN 9 (or 10) and
774 that the Internet is not available on VLAN 9 (or 10).
776 Q: Can I configure an IP address on a VLAN?
778 A: Yes. Use an "internal port" configured as an access port. For
779 example, the following configures IP address 192.168.0.7 on VLAN 9.
780 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
781 they have an 802.1Q header with VLAN 9. Conversely, traffic
782 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
786 ovs-vsctl add-port br0 eth0
787 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
788 ifconfig vlan9 192.168.0.7
790 Q: My OpenFlow controller doesn't see the VLANs that I expect.
792 A: The configuration for VLANs in the Open vSwitch database (e.g. via
793 ovs-vsctl) only affects traffic that goes through Open vSwitch's
794 implementation of the OpenFlow "normal switching" action. By
795 default, when Open vSwitch isn't connected to a controller and
796 nothing has been manually configured in the flow table, all traffic
797 goes through the "normal switching" action. But, if you set up
798 OpenFlow flows on your own, through a controller or using ovs-ofctl
799 or through other means, then you have to implement VLAN handling
802 You can use "normal switching" as a component of your OpenFlow
803 actions, e.g. by putting "normal" into the lists of actions on
804 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
805 controller. In situations where this is not suitable, you can
806 implement VLAN handling yourself, e.g.:
808 - If a packet comes in on an access port, and the flow table
809 needs to send it out on a trunk port, then the flow can add
810 the appropriate VLAN tag with the "mod_vlan_vid" action.
812 - If a packet comes in on a trunk port, and the flow table
813 needs to send it out on an access port, then the flow can
814 strip the VLAN tag with the "strip_vlan" action.
816 Q: I configured ports on a bridge as access ports with different VLAN
820 ovs-vsctl set-controller br0 tcp:192.168.0.10:6633
821 ovs-vsctl add-port br0 eth0
822 ovs-vsctl add-port br0 tap0 tag=9
823 ovs-vsctl add-port br0 tap1 tag=10
825 but the VMs running behind tap0 and tap1 can still communicate,
826 that is, they are not isolated from each other even though they are
829 A: Do you have a controller configured on br0 (as the commands above
830 do)? If so, then this is a variant on the previous question, "My
831 OpenFlow controller doesn't see the VLANs that I expect," and you
832 can refer to the answer there for more information.
838 Q: What versions of OpenFlow does Open vSwitch support?
840 A: Open vSwitch 1.9 and earlier support only OpenFlow 1.0 (plus
841 extensions that bring in many of the features from later versions
844 Open vSwitch versions 1.10 and later will have experimental support
845 for OpenFlow 1.2 and 1.3. On these versions of Open vSwitch, the
846 following command enables OpenFlow 1.0, 1.2, and 1.3 on bridge br0:
848 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow12,OpenFlow13
850 Support for OpenFlow 1.1 is incomplete enough that it cannot yet be
851 enabled, even experimentally.
853 Support for OpenFlow 1.2 and 1.3 is still incomplete. Work to be
854 done is tracked in OPENFLOW-1.1+ in the Open vSwitch source tree
855 (also via http://openvswitch.org/development/openflow-1-x-plan/).
856 When support for a given OpenFlow version is solidly implemented,
857 Open vSwitch will enable that version by default.
859 Q: I'm getting "error type 45250 code 0". What's that?
861 A: This is a Open vSwitch extension to OpenFlow error codes. Open
862 vSwitch uses this extension when it must report an error to an
863 OpenFlow controller but no standard OpenFlow error code is
866 Open vSwitch logs the errors that it sends to controllers, so the
867 easiest thing to do is probably to look at the ovs-vswitchd log to
868 find out what the error was.
870 If you want to dissect the extended error message yourself, the
871 format is documented in include/openflow/nicira-ext.h in the Open
872 vSwitch source distribution. The extended error codes are
873 documented in lib/ofp-errors.h.
875 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
876 doesn't actually appear through the OpenFlow connection, even
877 though I know that it's going through.
878 Q2: Some of the OpenFlow flows that my controller sets up don't seem
879 to apply to certain traffic, especially traffic between OVS and
880 the controller itself.
882 A: By default, Open vSwitch assumes that OpenFlow controllers are
883 connected "in-band", that is, that the controllers are actually
884 part of the network that is being controlled. In in-band mode,
885 Open vSwitch sets up special "hidden" flows to make sure that
886 traffic can make it back and forth between OVS and the controllers.
887 These hidden flows are higher priority than any flows that can be
888 set up through OpenFlow, and they are not visible through normal
889 OpenFlow flow table dumps.
891 Usually, the hidden flows are desirable and helpful, but
892 occasionally they can cause unexpected behavior. You can view the
893 full OpenFlow flow table, including hidden flows, on bridge br0
896 ovs-appctl bridge/dump-flows br0
898 to help you debug. The hidden flows are those with priorities
899 greater than 65535 (the maximum priority that can be set with
902 The DESIGN file at the top level of the Open vSwitch source
903 distribution describes the in-band model in detail.
905 If your controllers are not actually in-band (e.g. they are on
906 localhost via 127.0.0.1, or on a separate network), then you should
907 configure your controllers in "out-of-band" mode. If you have one
908 controller on bridge br0, then you can configure out-of-band mode
911 ovs-vsctl set controller br0 connection-mode=out-of-band
913 Q: I configured all my controllers for out-of-band control mode but
914 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
916 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
917 set-manager"). By default, Open vSwitch assumes that managers need
918 in-band rules set up on every bridge. You can disable these rules
921 ovs-vsctl set bridge br0 other-config:disable-in-band=true
923 This actually disables in-band control entirely for the bridge, as
924 if all the bridge's controllers were configured for out-of-band
927 Q: My OpenFlow controller doesn't see the VLANs that I expect.
929 A: See answer under "VLANs", above.
931 Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
932 but I got a funny message like this:
934 ofp_util|INFO|normalization changed ofp_match, details:
935 ofp_util|INFO| pre: nw_dst=192.168.0.1
938 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
939 match had disappeared, so that the flow ends up matching every
942 A: The term "normalization" in the log message means that a flow
943 cannot match on an L3 field without saying what L3 protocol is in
944 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
945 so the L3 field match was dropped.
947 In this case, the L3 protocol could be IP or ARP. A correct
948 command for each possibility is, respectively:
950 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
954 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
956 Similarly, a flow cannot match on an L4 field without saying what
957 L4 protocol is in use. For example, the flow match "tp_src=1234"
958 is, by itself, meaningless and will be ignored. Instead, to match
959 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
960 source port 1234, write "udp,tp_src=1234".
962 Q: How can I figure out the OpenFlow port number for a given port?
964 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
965 respond with an OFPT_FEATURES_REPLY that, among other information,
966 includes a mapping between OpenFlow port names and numbers. From a
967 command prompt, "ovs-ofctl show br0" makes such a request and
968 prints the response for switch br0.
970 The Interface table in the Open vSwitch database also maps OpenFlow
971 port names to numbers. To print the OpenFlow port number
972 associated with interface eth0, run:
974 ovs-vsctl get Interface eth0 ofport
976 You can print the entire mapping with:
978 ovs-vsctl -- --columns=name,ofport list Interface
980 but the output mixes together interfaces from all bridges in the
981 database, so it may be confusing if more than one bridge exists.
983 In the Open vSwitch database, ofport value -1 means that the
984 interface could not be created due to an error. (The Open vSwitch
985 log should indicate the reason.) ofport value [] (the empty set)
986 means that the interface hasn't been created yet. The latter is
987 normally an intermittent condition (unless ovs-vswitchd is not
990 Q: I added some flows with my controller or with ovs-ofctl, but when I
991 run "ovs-dpctl dump-flows" I don't see them.
993 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
994 won't display the information that you want. You want to use
995 "ovs-ofctl dump-flows" instead.
997 Q: It looks like each of the interfaces in my bonded port shows up
998 as an individual OpenFlow port. Is that right?
1000 A: Yes, Open vSwitch makes individual bond interfaces visible as
1001 OpenFlow ports, rather than the bond as a whole. The interfaces
1002 are treated together as a bond for only a few purposes:
1004 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1005 controller is not configured, this happens implicitly to
1008 - Mirrors configured for output to a bonded port.
1010 It would make a lot of sense for Open vSwitch to present a bond as
1011 a single OpenFlow port. If you want to contribute an
1012 implementation of such a feature, please bring it up on the Open
1013 vSwitch development mailing list at dev@openvswitch.org.
1015 Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1016 multiple hosts, and other components. The behavior isn't what I
1019 A: To debug network behavior problems, trace the path of a packet,
1020 hop-by-hop, from its origin in one host to a remote host. If
1021 that's correct, then trace the path of the response packet back to
1024 Usually a simple ICMP echo request and reply ("ping") packet is
1025 good enough. Start by initiating an ongoing "ping" from the origin
1026 host to a remote host. If you are tracking down a connectivity
1027 problem, the "ping" will not display any successful output, but
1028 packets are still being sent. (In this case the packets being sent
1029 are likely ARP rather than ICMP.)
1031 Tools available for tracing include the following:
1033 - "tcpdump" and "wireshark" for observing hops across network
1034 devices, such as Open vSwitch internal devices and physical
1037 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1038 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1039 These tools allow one to observe the actions being taken on
1040 packets in ongoing flows.
1042 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1043 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1044 documentation, and "Why are there so many different ways to
1045 dump flows?" above for some background.
1047 - "ovs-appctl ofproto/trace" to observe the logic behind how
1048 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1049 documentation. You can out more details about a given flow
1050 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1051 a flow from the output into an "ovs-appctl ofproto/trace"
1054 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1055 observe what goes on at these physical hops.
1057 Starting at the origin of a given packet, observe the packet at
1058 each hop in turn. For example, in one plausible scenario, you
1061 1. "tcpdump" the "eth" interface through which an ARP egresses
1062 a VM, from inside the VM.
1064 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1065 ingresses the host machine.
1067 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1068 the host interface through which the ARP egresses the
1069 physical machine. You may need to use "ovs-dpctl show" to
1070 interpret the port numbers. If the output seems surprising,
1071 you can use "ovs-appctl ofproto/trace" to observe details of
1072 how ovs-vswitchd determined the actions in the "ovs-dpctl
1075 4. "tcpdump" the "eth" interface through which the ARP egresses
1076 the physical machine.
1078 5. "tcpdump" the "eth" interface through which the ARP
1079 ingresses the physical machine, at the remote host that
1082 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1083 remote host that receives the ARP and observe the VM "vif"
1084 or "tap" interface to which the flow is directed. Again,
1085 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1087 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1090 8. "tcpdump" the "eth" interface through which the ARP
1091 ingresses a VM, from inside the VM.
1093 It is likely that during one of these steps you will figure out the
1094 problem. If not, then follow the ARP reply back to the origin, in
1100 bugs@openvswitch.org
1101 http://openvswitch.org/