1 Frequently Asked Questions
2 ==========================
4 Open vSwitch <http://openvswitch.org>
9 ### Q: What is Open vSwitch?
11 A: Open vSwitch is a production quality open source software switch
12 designed to be used as a vswitch in virtualized server
13 environments. A vswitch forwards traffic between different VMs on
14 the same physical host and also forwards traffic between VMs and
15 the physical network. Open vSwitch supports standard management
16 interfaces (e.g. sFlow, NetFlow, IPFIX, RSPAN, CLI), and is open to
17 programmatic extension and control using OpenFlow and the OVSDB
20 Open vSwitch as designed to be compatible with modern switching
21 chipsets. This means that it can be ported to existing high-fanout
22 switches allowing the same flexible control of the physical
23 infrastructure as the virtual infrastructure. It also means that
24 Open vSwitch will be able to take advantage of on-NIC switching
25 chipsets as their functionality matures.
27 ### Q: What virtualization platforms can use Open vSwitch?
29 A: Open vSwitch can currently run on any Linux-based virtualization
30 platform (kernel 2.6.32 and newer), including: KVM, VirtualBox, Xen,
31 Xen Cloud Platform, XenServer. As of Linux 3.3 it is part of the
32 mainline kernel. The bulk of the code is written in platform-
33 independent C and is easily ported to other environments. We welcome
34 inquires about integrating Open vSwitch with other virtualization
37 ### Q: How can I try Open vSwitch?
39 A: The Open vSwitch source code can be built on a Linux system. You can
40 build and experiment with Open vSwitch on any Linux machine.
41 Packages for various Linux distributions are available on many
42 platforms, including: Debian, Ubuntu, Fedora.
44 You may also download and run a virtualization platform that already
45 has Open vSwitch integrated. For example, download a recent ISO for
46 XenServer or Xen Cloud Platform. Be aware that the version
47 integrated with a particular platform may not be the most recent Open
50 ### Q: Does Open vSwitch only work on Linux?
52 A: No, Open vSwitch has been ported to a number of different operating
53 systems and hardware platforms. Most of the development work occurs
54 on Linux, but the code should be portable to any POSIX system. We've
55 seen Open vSwitch ported to a number of different platforms,
56 including FreeBSD, Windows, and even non-POSIX embedded systems.
58 By definition, the Open vSwitch Linux kernel module only works on
59 Linux and will provide the highest performance. However, a userspace
60 datapath is available that should be very portable.
62 ### Q: What's involved with porting Open vSwitch to a new platform or switching ASIC?
64 A: The [PORTING.md] document describes how one would go about
65 porting Open 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
72 [WHY-OVS.md] for a more detailed description of how Open vSwitch
73 relates to the Linux Bridge.
75 ### Q: How is Open vSwitch related to distributed virtual switches like the VMware vNetwork distributed switch or the Cisco Nexus 1000V?
77 A: Distributed vswitch applications (e.g., VMware vNetwork distributed
78 switch, Cisco Nexus 1000V) provide a centralized way to configure and
79 monitor the network state of VMs that are spread across many physical
80 hosts. Open vSwitch is not a distributed vswitch itself, rather it
81 runs on each physical host and supports remote management in a way
82 that makes it easier for developers of virtualization/cloud
83 management platforms to offer distributed vswitch capabilities.
85 To aid in distribution, Open vSwitch provides two open protocols that
86 are specially designed for remote management in virtualized network
87 environments: OpenFlow, which exposes flow-based forwarding state,
88 and the OVSDB management protocol, which exposes switch port state.
89 In addition to the switch implementation itself, Open vSwitch
90 includes tools (ovs-ofctl, ovs-vsctl) that developers can script and
91 extend to provide distributed vswitch capabilities that are closely
92 integrated with their virtualization management platform.
94 ### Q: Why doesn't Open vSwitch support distribution?
96 A: Open vSwitch is intended to be a useful component for building
97 flexible network infrastructure. There are many different approaches
98 to distribution which balance trade-offs between simplicity,
99 scalability, hardware compatibility, convergence times, logical
100 forwarding model, etc. The goal of Open vSwitch is to be able to
101 support all as a primitive building block rather than choose a
102 particular point in the distributed design space.
104 ### Q: How can I contribute to the Open vSwitch Community?
106 A: You can start by joining the mailing lists and helping to answer
107 questions. You can also suggest improvements to documentation. If
108 you have a feature or bug you would like to work on, send a mail to
109 one of the mailing lists:
111 http://openvswitch.org/mlists/
113 ### Q: Why can I no longer connect to my OpenFlow controller or OVSDB manager?
115 A: Starting in OVS 2.4, we switched the default ports to the
116 IANA-specified port numbers for OpenFlow (6633->6653) and OVSDB
117 (6632->6640). We recommend using these port numbers, but if you
118 cannot, all the programs allow overriding the default port. See the
119 appropriate man page.
125 ### Q: What does it mean for an Open vSwitch release to be LTS (long-term support)?
127 A: All official releases have been through a comprehensive testing
128 process and are suitable for production use. Planned releases will
129 occur several times a year. If a significant bug is identified in an
130 LTS release, we will provide an updated release that includes the
131 fix. Releases that are not LTS may not be fixed and may just be
132 supplanted by the next major release. The current LTS release is
135 ### Q: What Linux kernel versions does each Open vSwitch release work with?
137 A: The following table lists the Linux kernel versions against which the
138 given versions of the Open vSwitch kernel module will successfully
139 build. The Linux kernel versions are upstream kernel versions, so
140 Linux kernels modified from the upstream sources may not build in
141 some cases even if they are based on a supported version. This is
142 most notably true of Red Hat Enterprise Linux (RHEL) kernels, which
143 are extensively modified from upstream.
145 | Open vSwitch | Linux kernel
146 |:------------:|:-------------:
147 | 1.4.x | 2.6.18 to 3.2
148 | 1.5.x | 2.6.18 to 3.2
149 | 1.6.x | 2.6.18 to 3.2
150 | 1.7.x | 2.6.18 to 3.3
151 | 1.8.x | 2.6.18 to 3.4
152 | 1.9.x | 2.6.18 to 3.8
153 | 1.10.x | 2.6.18 to 3.8
154 | 1.11.x | 2.6.18 to 3.8
155 | 2.0.x | 2.6.32 to 3.10
156 | 2.1.x | 2.6.32 to 3.11
157 | 2.3.x | 2.6.32 to 3.14
158 | 2.4.x | 2.6.32 to 4.0
159 | 2.5.x | 2.6.32 to 4.2
161 Open vSwitch userspace should also work with the Linux kernel module
162 built into Linux 3.3 and later.
164 Open vSwitch userspace is not sensitive to the Linux kernel version.
165 It should build against almost any kernel, certainly against 2.6.32
168 ### Q: I get an error like this when I configure Open vSwitch:
170 configure: error: Linux kernel in <dir> is version <x>, but
171 version newer than <y> is not supported (please refer to the
176 A: You have the following options:
178 - Use the Linux kernel module supplied with the kernel that you are
179 using. (See also the following FAQ.)
181 - If there is a newer released version of Open vSwitch, consider
182 building that one, because it may support the kernel that you are
183 building against. (To find out, consult the table in the
186 - The Open vSwitch "master" branch may support the kernel that you
187 are using, so consider building the kernel module from "master".
189 All versions of Open vSwitch userspace are compatible with all
190 versions of the Open vSwitch kernel module, so you do not have to
191 use the kernel module from one source along with the userspace
192 programs from the same source.
194 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
196 A: The kernel module in upstream Linux does not include support for
197 LISP. Work is in progress to add support for LISP to the upstream
198 Linux version of the Open vSwitch kernel module. For now, if you
199 need this feature, use the kernel module from the Open vSwitch
200 distribution instead of the upstream Linux kernel module.
202 Certain features require kernel support to function or to have
203 reasonable performance. If the ovs-vswitchd log file indicates that
204 a feature is not supported, consider upgrading to a newer upstream
205 Linux release or using the kernel module paired with the userspace
208 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
210 A: Support for tunnels was added to the upstream Linux kernel module
211 after the rest of Open vSwitch. As a result, some kernels may contain
212 support for Open vSwitch but not tunnels. The minimum kernel version
213 that supports each tunnel protocol is:
215 | Protocol | Linux Kernel
216 |:--------:|:-------------:
220 | LISP | <not upstream>
221 | STT | <not upstream>
223 If you are using a version of the kernel that is older than the one
224 listed above, it is still possible to use that tunnel protocol. However,
225 you must compile and install the kernel module included with the Open
226 vSwitch distribution rather than the one on your machine. If problems
227 persist after doing this, check to make sure that the module that is
228 loaded is the one you expect.
230 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
232 A: Generating outer UDP checksums requires kernel support that was not
233 part of the initial implementation of these protocols. If using the
234 upstream Linux Open vSwitch module, you must use kernel 4.0 or
235 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
236 support UDP checksums.
238 ### Q: What features are not available when using the userspace datapath?
240 A: Tunnel virtual ports are not supported, as described in the
241 previous answer. It is also not possible to use queue-related
242 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
243 may not be transmitted.
245 ### Q: What Linux kernel versions does IPFIX flow monitoring work with?
247 A: IPFIX flow monitoring requires the Linux kernel module from Linux
248 3.10 or later, or the out-of-tree module from Open vSwitch version
251 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
253 In general, the Open vSwitch userspace should be used with the
254 kernel version included in the same release or with the version
255 from upstream Linux. However, when upgrading between two releases
256 of Open vSwitch it is best to migrate userspace first to reduce
257 the possibility of incompatibilities.
259 ### Q: What happened to the bridge compatibility feature?
261 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
262 When it was enabled, Open vSwitch imitated the interface of the
263 Linux kernel "bridge" module. This allowed users to drop Open
264 vSwitch into environments designed to use the Linux kernel bridge
265 module without adapting the environment to use Open vSwitch.
267 Open vSwitch 1.10 and later do not support bridge compatibility.
268 The feature was dropped because version 1.10 adopted a new internal
269 architecture that made bridge compatibility difficult to maintain.
270 Now that many environments use OVS directly, it would be rarely
273 To use bridge compatibility, install OVS 1.9 or earlier, including
274 the accompanying kernel modules (both the main and bridge
275 compatibility modules), following the instructions that come with
276 the release. Be sure to start the ovs-brcompatd daemon.
282 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
284 A: In networking, the terms "bridge" and "switch" are synonyms. Open
285 vSwitch implements an Ethernet switch, which means that it is also
288 ### Q: What's a VLAN?
290 A: See the "VLAN" section below.
296 ### Q: How do I configure a port as an access port?
298 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
299 the following commands configure br0 with eth0 as a trunk port (the
300 default) and tap0 as an access port for VLAN 9:
303 ovs-vsctl add-port br0 eth0
304 ovs-vsctl add-port br0 tap0 tag=9
306 If you want to configure an already added port as an access port,
307 use "ovs-vsctl set", e.g.:
309 ovs-vsctl set port tap0 tag=9
311 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
313 A: The following commands configure br0 with eth0 and tap0 as trunk
314 ports. All traffic coming in or going out on eth0 or tap0 is also
315 mirrored to tap1; any traffic arriving on tap1 is dropped:
318 ovs-vsctl add-port br0 eth0
319 ovs-vsctl add-port br0 tap0
320 ovs-vsctl add-port br0 tap1 \
321 -- --id=@p get port tap1 \
322 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
323 -- set bridge br0 mirrors=@m
325 To later disable mirroring, run:
327 ovs-vsctl clear bridge br0 mirrors
329 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
331 A: Yes. How you configure it depends on what you mean by "promiscuous
334 - Conventionally, "promiscuous mode" is a feature of a network
335 interface card. Ordinarily, a NIC passes to the CPU only the
336 packets actually destined to its host machine. It discards
337 the rest to avoid wasting memory and CPU cycles. When
338 promiscuous mode is enabled, however, it passes every packet
339 to the CPU. On an old-style shared-media or hub-based
340 network, this allows the host to spy on all packets on the
341 network. But in the switched networks that are almost
342 everywhere these days, promiscuous mode doesn't have much
343 effect, because few packets not destined to a host are
344 delivered to the host's NIC.
346 This form of promiscuous mode is configured in the guest OS of
347 the VMs on your bridge, e.g. with "ifconfig".
349 - The VMware vSwitch uses a different definition of "promiscuous
350 mode". When you configure promiscuous mode on a VMware vNIC,
351 the vSwitch sends a copy of every packet received by the
352 vSwitch to that vNIC. That has a much bigger effect than just
353 enabling promiscuous mode in a guest OS. Rather than getting
354 a few stray packets for which the switch does not yet know the
355 correct destination, the vNIC gets every packet. The effect
356 is similar to replacing the vSwitch by a virtual hub.
358 This "promiscuous mode" is what switches normally call "port
359 mirroring" or "SPAN". For information on how to configure
360 SPAN, see "How do I configure a port as a SPAN port, that is,
361 enable mirroring of all traffic to that port?"
363 ### Q: How do I configure a DPDK port as an access port?
365 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
367 If your version is DPDK-enabled it will support the --dpdk
368 argument on the command line and will display lines with
369 "EAL:..." during startup when --dpdk is supplied.
371 Secondly, when adding a DPDK port, unlike a system port, the
372 type for the interface must be specified. For example;
375 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
377 Finally, it is required that DPDK port names begin with 'dpdk'.
379 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
382 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
384 A: The following commands configure br0 with eth0 as a trunk port and
385 tap0 as an access port for VLAN 10. All traffic coming in or going
386 out on tap0, as well as traffic coming in or going out on eth0 in
387 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
388 VLAN 10, in cases where one is present, is dropped as part of
392 ovs-vsctl add-port br0 eth0
393 ovs-vsctl add-port br0 tap0 tag=10
395 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
397 -- set bridge br0 mirrors=@m
399 To later disable mirroring, run:
401 ovs-vsctl clear bridge br0 mirrors
403 Mirroring to a VLAN can disrupt a network that contains unmanaged
404 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
405 GRE tunnel has fewer caveats than mirroring to a VLAN and should
406 generally be preferred.
408 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
410 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
411 of the specified output-vlan. This loss of information may make
412 the mirrored traffic too hard to interpret.
414 To mirror multiple VLANs, use the commands above, but specify a
415 comma-separated list of VLANs as the value for select-vlan. To
416 mirror every VLAN, use the commands above, but omit select-vlan and
419 When a packet arrives on a VLAN that is used as a mirror output
420 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
421 floods the packet across all the ports for which the mirror output
422 VLAN is configured. (If an OpenFlow controller is in use, then it
423 can override this behavior through the flow table.) If OVS is used
424 as an intermediate switch, rather than an edge switch, this ensures
425 that the RSPAN traffic is distributed through the network.
427 Mirroring to a VLAN can disrupt a network that contains unmanaged
428 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
429 GRE tunnel has fewer caveats than mirroring to a VLAN and should
430 generally be preferred.
432 ### Q: How do I configure mirroring of all traffic to a GRE tunnel?
434 A: The following commands configure br0 with eth0 and tap0 as trunk
435 ports. All traffic coming in or going out on eth0 or tap0 is also
436 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
437 traffic arriving on gre0 is dropped:
440 ovs-vsctl add-port br0 eth0
441 ovs-vsctl add-port br0 tap0
442 ovs-vsctl add-port br0 gre0 \
443 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
444 -- --id=@p get port gre0 \
445 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
446 -- set bridge br0 mirrors=@m
448 To later disable mirroring and destroy the GRE tunnel:
450 ovs-vsctl clear bridge br0 mirrors
451 ovs-vcstl del-port br0 gre0
453 ### Q: Does Open vSwitch support ERSPAN?
455 A: No. ERSPAN is an undocumented proprietary protocol. As an
456 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
459 ### Q: How do I connect two bridges?
461 A: First, why do you want to do this? Two connected bridges are not
462 much different from a single bridge, so you might as well just have
463 a single bridge with all your ports on it.
465 If you still want to connect two bridges, you can use a pair of
466 patch ports. The following example creates bridges br0 and br1,
467 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
468 and br1 with a pair of patch ports.
471 ovs-vsctl add-port br0 eth0
472 ovs-vsctl add-port br0 tap0
474 ovs-vsctl add-port br1 tap1
476 -- add-port br0 patch0 \
477 -- set interface patch0 type=patch options:peer=patch1 \
478 -- add-port br1 patch1 \
479 -- set interface patch1 type=patch options:peer=patch0
481 Bridges connected with patch ports are much like a single bridge.
482 For instance, if the example above also added eth1 to br1, and both
483 eth0 and eth1 happened to be connected to the same next-hop switch,
484 then you could loop your network just as you would if you added
485 eth0 and eth1 to the same bridge (see the "Configuration Problems"
486 section below for more information).
488 If you are using Open vSwitch 1.9 or an earlier version, then you
489 need to be using the kernel module bundled with Open vSwitch rather
490 than the one that is integrated into Linux 3.3 and later, because
491 Open vSwitch 1.9 and earlier versions need kernel support for patch
492 ports. This also means that in Open vSwitch 1.9 and earlier, patch
493 ports will not work with the userspace datapath, only with the
496 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
498 A: Open vSwitch does not support such a configuration.
499 Bridges always have their local ports.
502 Implementation Details
503 ----------------------
505 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
507 A: Open vSwitch uses different kinds of flows for different purposes:
509 - OpenFlow flows are the most important kind of flow. OpenFlow
510 controllers use these flows to define a switch's policy.
511 OpenFlow flows support wildcards, priorities, and multiple
514 When in-band control is in use, Open vSwitch sets up a few
515 "hidden" flows, with priority higher than a controller or the
516 user can configure, that are not visible via OpenFlow. (See
517 the "Controller" section of the FAQ for more information
520 - The Open vSwitch software switch implementation uses a second
521 kind of flow internally. These flows, called "datapath" or
522 "kernel" flows, do not support priorities and comprise only a
523 single table, which makes them suitable for caching. (Like
524 OpenFlow flows, datapath flows do support wildcarding, in Open
525 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
526 also support different actions and number ports differently.
528 Datapath flows are an implementation detail that is subject to
529 change in future versions of Open vSwitch. Even with the
530 current version of Open vSwitch, hardware switch
531 implementations do not necessarily use this architecture.
533 Users and controllers directly control only the OpenFlow flow
534 table. Open vSwitch manages the datapath flow table itself, so
535 users should not normally be concerned with it.
537 ### Q: Why are there so many different ways to dump flows?
539 A: Open vSwitch has two kinds of flows (see the previous question), so
540 it has commands with different purposes for dumping each kind of
543 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
544 hidden flows. This is the most commonly useful form of flow
545 dump. (Unlike the other commands, this should work with any
546 OpenFlow switch, not just Open vSwitch.)
548 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
549 including hidden flows. This is occasionally useful for
550 troubleshooting suspected issues with in-band control.
552 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
553 entries for a Linux kernel-based datapath. In Open vSwitch
554 1.10 and later, ovs-vswitchd merges multiple switches into a
555 single datapath, so it will show all the flows on all your
556 kernel-based switches. This command can occasionally be
557 useful for debugging.
559 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
560 dumps datapath flows for only the specified bridge, regardless
563 ### Q: How does multicast snooping works with VLANs?
565 A: Open vSwitch maintains snooping tables for each VLAN.
567 ### Q: Can OVS populate the kernel flow table in advance instead of in reaction to packets?
569 A: No. There are several reasons:
571 - Kernel flows are not as sophisticated as OpenFlow flows, which
572 means that some OpenFlow policies could require a large number of
573 kernel flows. The "conjunctive match" feature is an extreme
574 example: the number of kernel flows it requires is the product of
575 the number of flows in each dimension.
577 - With multiple OpenFlow flow tables and simple sets of actions, the
578 number of kernel flows required can be as large as the product of
579 the number of flows in each dimension. With more sophisticated
580 actions, the number of kernel flows could be even larger.
582 - Open vSwitch is designed so that any version of OVS userspace
583 interoperates with any version of the OVS kernel module. This
584 forward and backward compatibility requires that userspace observe
585 how the kernel module parses received packets. This is only
586 possible in a straightforward way when userspace adds kernel flows
587 in reaction to received packets.
589 For more relevant information on the architecture of Open vSwitch,
590 please read "The Design and Implementation of Open vSwitch",
591 published in USENIX NSDI 2015.
597 ### Q: I just upgraded and I see a performance drop. Why?
599 A: The OVS kernel datapath may have been updated to a newer version than
600 the OVS userspace components. Sometimes new versions of OVS kernel
601 module add functionality that is backwards compatible with older
602 userspace components but may cause a drop in performance with them.
603 Especially, if a kernel module from OVS 2.1 or newer is paired with
604 OVS userspace 1.10 or older, there will be a performance drop for
607 Updating the OVS userspace components to the latest released
608 version should fix the performance degradation.
610 To get the best possible performance and functionality, it is
611 recommended to pair the same versions of the kernel module and OVS
615 Configuration Problems
616 ----------------------
618 ### Q: I created a bridge and added my Ethernet port to it, using commands
622 ovs-vsctl add-port br0 eth0
624 and as soon as I ran the "add-port" command I lost all connectivity
627 A: A physical Ethernet device that is part of an Open vSwitch bridge
628 should not have an IP address. If one does, then that IP address
629 will not be fully functional.
631 You can restore functionality by moving the IP address to an Open
632 vSwitch "internal" device, such as the network device named after
633 the bridge itself. For example, assuming that eth0's IP address is
634 192.168.128.5, you could run the commands below to fix up the
637 ifconfig eth0 0.0.0.0
638 ifconfig br0 192.168.128.5
640 (If your only connection to the machine running OVS is through the
641 IP address in question, then you would want to run all of these
642 commands on a single command line, or put them into a script.) If
643 there were any additional routes assigned to eth0, then you would
644 also want to use commands to adjust these routes to go through br0.
646 If you use DHCP to obtain an IP address, then you should kill the
647 DHCP client that was listening on the physical Ethernet interface
648 (e.g. eth0) and start one listening on the internal interface
649 (e.g. br0). You might still need to manually clear the IP address
650 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
652 There is no compelling reason why Open vSwitch must work this way.
653 However, this is the way that the Linux kernel bridge module has
654 always worked, so it's a model that those accustomed to Linux
655 bridging are already used to. Also, the model that most people
656 expect is not implementable without kernel changes on all the
657 versions of Linux that Open vSwitch supports.
659 By the way, this issue is not specific to physical Ethernet
660 devices. It applies to all network devices except Open vSwitch
663 ### Q: I created a bridge and added a couple of Ethernet ports to it,
664 ### using commands like these:
667 ovs-vsctl add-port br0 eth0
668 ovs-vsctl add-port br0 eth1
670 and now my network seems to have melted: connectivity is unreliable
671 (even connectivity that doesn't go through Open vSwitch), all the
672 LEDs on my physical switches are blinking, wireshark shows
673 duplicated packets, and CPU usage is very high.
675 A: More than likely, you've looped your network. Probably, eth0 and
676 eth1 are connected to the same physical Ethernet switch. This
677 yields a scenario where OVS receives a broadcast packet on eth0 and
678 sends it out on eth1, then the physical switch connected to eth1
679 sends the packet back on eth0, and so on forever. More complicated
680 scenarios, involving a loop through multiple switches, are possible
683 The solution depends on what you are trying to do:
685 - If you added eth0 and eth1 to get higher bandwidth or higher
686 reliability between OVS and your physical Ethernet switch,
687 use a bond. The following commands create br0 and then add
688 eth0 and eth1 as a bond:
691 ovs-vsctl add-bond br0 bond0 eth0 eth1
693 Bonds have tons of configuration options. Please read the
694 documentation on the Port table in ovs-vswitchd.conf.db(5)
697 Configuration for DPDK-enabled interfaces is slightly less
698 straightforward: see [INSTALL.DPDK.md].
700 - Perhaps you don't actually need eth0 and eth1 to be on the
701 same bridge. For example, if you simply want to be able to
702 connect each of them to virtual machines, then you can put
703 each of them on a bridge of its own:
706 ovs-vsctl add-port br0 eth0
709 ovs-vsctl add-port br1 eth1
711 and then connect VMs to br0 and br1. (A potential
712 disadvantage is that traffic cannot directly pass between br0
713 and br1. Instead, it will go out eth0 and come back in eth1,
716 - If you have a redundant or complex network topology and you
717 want to prevent loops, turn on spanning tree protocol (STP).
718 The following commands create br0, enable STP, and add eth0
719 and eth1 to the bridge. The order is important because you
720 don't want have to have a loop in your network even
724 ovs-vsctl set bridge br0 stp_enable=true
725 ovs-vsctl add-port br0 eth0
726 ovs-vsctl add-port br0 eth1
728 The Open vSwitch implementation of STP is not well tested.
729 Please report any bugs you observe, but if you'd rather avoid
730 acting as a beta tester then another option might be your
733 ### Q: I can't seem to use Open vSwitch in a wireless network.
735 A: Wireless base stations generally only allow packets with the source
736 MAC address of NIC that completed the initial handshake.
737 Therefore, without MAC rewriting, only a single device can
738 communicate over a single wireless link.
740 This isn't specific to Open vSwitch, it's enforced by the access
741 point, so the same problems will show up with the Linux bridge or
742 any other way to do bridging.
744 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
746 A: PPP most commonly carries IP packets, but Open vSwitch works only
747 with Ethernet frames. The correct way to interface PPP to an
748 Ethernet network is usually to use routing instead of switching.
750 ### Q: Is there any documentation on the database tables and fields?
752 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
754 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
755 I only see a datapath called "ovs-system". How can I see datapath
756 information about a particular bridge?
758 A: In version 1.9.0, OVS switched to using a single datapath that is
759 shared by all bridges of that type. The "ovs-appctl dpif/*"
760 commands provide similar functionality that is scoped by the bridge.
762 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
763 see the port in the datapath?
765 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
766 cannot be loaded at the same time. It is likely that on your system the
767 Linux GRE module is already loaded and blocking OVS (to confirm, check
768 dmesg for errors regarding GRE registration). To fix this, unload all
769 GRE modules that appear in lsmod as well as the OVS kernel module. You
770 can then reload the OVS module following the directions in
771 [INSTALL.md], which will ensure that dependencies are satisfied.
773 ### Q: Open vSwitch does not seem to obey my packet filter rules.
775 A: It depends on mechanisms and configurations you want to use.
777 You cannot usefully use typical packet filters, like iptables, on
778 physical Ethernet ports that you add to an Open vSwitch bridge.
779 This is because Open vSwitch captures packets from the interface at
780 a layer lower below where typical packet-filter implementations
781 install their hooks. (This actually applies to any interface of
782 type "system" that you might add to an Open vSwitch bridge.)
784 You can usefully use typical packet filters on Open vSwitch
785 internal ports as they are mostly ordinary interfaces from the point
786 of view of packet filters.
788 For example, suppose you create a bridge br0 and add Ethernet port
789 eth0 to it. Then you can usefully add iptables rules to affect the
790 internal interface br0, but not the physical interface eth0. (br0
791 is also where you would add an IP address, as discussed elsewhere
794 For simple filtering rules, it might be possible to achieve similar
795 results by installing appropriate OpenFlow flows instead.
797 If the use of a particular packet filter setup is essential, Open
798 vSwitch might not be the best choice for you. On Linux, you might
799 want to consider using the Linux Bridge. (This is the only choice if
800 you want to use ebtables rules.) On NetBSD, you might want to
801 consider using the bridge(4) with BRIDGE_IPF option.
803 ### Q: It seems that Open vSwitch does nothing when I removed a port and
804 then immediately put it back. For example, consider that p1 is
805 a port of type=internal:
807 ovs-vsctl del-port br0 p1 -- \
809 set interface p1 type=internal
811 A: It's an expected behaviour.
813 If del-port and add-port happen in a single OVSDB transaction as
814 your example, Open vSwitch always "skips" the intermediate steps.
815 Even if they are done in multiple transactions, it's still allowed
816 for Open vSwitch to skip the intermediate steps and just implement
817 the overall effect. In both cases, your example would be turned
820 If you want to make Open vSwitch actually destroy and then re-create
821 the port for some side effects like resetting kernel setting for the
822 corresponding interface, you need to separate operations into multiple
823 OVSDB transactions and ensure that at least the first one does not have
824 --no-wait. In the following example, the first ovs-vsctl will block
825 until Open vSwitch reloads the new configuration and removes the port:
827 ovs-vsctl del-port br0 p1
828 ovs-vsctl add-port br0 p1 -- \
829 set interface p1 type=internal
831 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
832 it takes too long (minutes or hours) to do it with ovs-vsctl. How
835 A: If you add them one at a time with ovs-vsctl, it can take a long
836 time to add thousands of ports to an Open vSwitch bridge. This is
837 because every invocation of ovs-vsctl first reads the current
838 configuration from OVSDB. As the number of ports grows, this
839 starts to take an appreciable amount of time, and when it is
840 repeated thousands of times the total time becomes significant.
842 The solution is to add the ports in one invocation of ovs-vsctl (or
843 a small number of them). For example, using bash:
846 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
849 takes seconds, not minutes or hours, in the OVS sandbox environment.
851 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
852 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
855 A: Open vSwitch wasn't able to create the bridge. Check the
856 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
857 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
859 In general, the Open vSwitch database reflects the desired
860 configuration state. ovs-vswitchd monitors the database and, when
861 it changes, reconfigures the system to reflect the new desired
862 state. This normally happens very quickly. Thus, a discrepancy
863 between the database and the actual state indicates that
864 ovs-vswitchd could not implement the configuration, and so one
865 should check the log to find out why. (Another possible cause is
866 that ovs-vswitchd is not running. This will make "ovs-vsctl"
867 commands hang, if they change the configuration, unless one
868 specifies "--no-wait".)
870 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
871 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
872 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
875 A: Open vSwitch wasn't able to create the port. Check the
876 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
877 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
878 see the previous question for more information.
880 You may want to upgrade to Open vSwitch 2.3 (or later), in which
881 ovs-vsctl will immediately report when there is an issue creating a
884 ### Q: I created a tap device tap0, configured an IP address on it, and
885 added it to a bridge, like this:
888 ifconfig tap0 192.168.0.123
890 ovs-vsctl add-port br0 tap0
892 I expected that I could then use this IP address to contact other
893 hosts on the network, but it doesn't work. Why not?
895 A: The short answer is that this is a misuse of a "tap" device. Use
896 an "internal" device implemented by Open vSwitch, which works
897 differently and is designed for this use. To solve this problem
898 with an internal device, instead run:
901 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
902 ifconfig int0 192.168.0.123
904 Even more simply, you can take advantage of the internal port that
905 every bridge has under the name of the bridge:
908 ifconfig br0 192.168.0.123
910 In more detail, a "tap" device is an interface between the Linux
911 (or *BSD) network stack and a user program that opens it as a
912 socket. When the "tap" device transmits a packet, it appears in
913 the socket opened by the userspace program. Conversely, when the
914 userspace program writes to the "tap" socket, the kernel TCP/IP
915 stack processes the packet as if it had been received by the "tap"
918 Consider the configuration above. Given this configuration, if you
919 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
920 routing stack will transmit an ARP on the tap0 device. Open
921 vSwitch userspace treats "tap" devices just like any other network
922 device; that is, it doesn't open them as "tap" sockets. That means
923 that the ARP packet will simply get dropped.
925 You might wonder why the Open vSwitch kernel module doesn't
926 intercept the ARP packet and bridge it. After all, Open vSwitch
927 intercepts packets on other devices. The answer is that Open
928 vSwitch only intercepts *received* packets, but this is a packet
929 being transmitted. The same thing happens for all other types of
930 network devices, except for Open vSwitch "internal" ports. If you,
931 for example, add a physical Ethernet port to an OVS bridge,
932 configure an IP address on a physical Ethernet port, and then issue
933 a "ping" to an address in that subnet, the same thing happens: an
934 ARP gets transmitted on the physical Ethernet port and Open vSwitch
935 never sees it. (You should not do that, as documented at the
936 beginning of this section.)
938 It can make sense to add a "tap" device to an Open vSwitch bridge,
939 if some userspace program (other than Open vSwitch) has opened the
940 tap socket. This is the case, for example, if the "tap" device was
941 created by KVM (or QEMU) to simulate a virtual NIC. In such a
942 case, when OVS bridges a packet to the "tap" device, the kernel
943 forwards that packet to KVM in userspace, which passes it along to
944 the VM, and in the other direction, when the VM sends a packet, KVM
945 writes it to the "tap" socket, which causes OVS to receive it and
946 bridge it to the other OVS ports. Please note that in such a case
947 no IP address is configured on the "tap" device (there is normally
948 an IP address configured in the virtual NIC inside the VM, but this
949 is not visible to the host Linux kernel or to Open vSwitch).
951 There is one special case in which Open vSwitch does directly read
952 and write "tap" sockets. This is an implementation detail of the
953 Open vSwitch userspace switch, which implements its "internal"
954 ports as Linux (or *BSD) "tap" sockets. In such a userspace
955 switch, OVS receives packets sent on the "tap" device used to
956 implement an "internal" port by reading the associated "tap"
957 socket, and bridges them to the rest of the switch. In the other
958 direction, OVS transmits packets bridged to the "internal" port by
959 writing them to the "tap" socket, causing them to be processed by
960 the kernel TCP/IP stack as if they had been received on the "tap"
961 device. Users should not need to be concerned with this
962 implementation detail.
964 Open vSwitch has a network device type called "tap". This is
965 intended only for implementing "internal" ports in the OVS
966 userspace switch and should not be used otherwise. In particular,
967 users should not configure KVM "tap" devices as type "tap" (use
968 type "system", the default, instead).
971 Quality of Service (QoS)
972 ------------------------
974 ### Q: Does OVS support Quality of Service (QoS)?
976 A: Yes. For traffic that egresses from a switch, OVS supports traffic
977 shaping; for traffic that ingresses into a switch, OVS support
978 policing. Policing is a simple form of quality-of-service that
979 simply drops packets received in excess of the configured rate. Due
980 to its simplicity, policing is usually less accurate and less
981 effective than egress traffic shaping, which queues packets.
983 Keep in mind that ingress and egress are from the perspective of the
984 switch. That means that egress shaping limits the rate at which
985 traffic is allowed to transmit from a physical interface, but the
986 rate at which traffic will be received on a virtual machine's VIF.
987 For ingress policing, the behavior is the opposite.
989 ### Q: How do I configure egress traffic shaping?
991 A: Suppose that you want to set up bridge br0 connected to physical
992 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
993 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
994 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
995 could configure the bridge this way:
999 add-port br0 eth0 -- \
1000 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
1001 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
1002 set port eth0 qos=@newqos -- \
1003 --id=@newqos create qos type=linux-htb \
1004 other-config:max-rate=1000000000 \
1005 queues:123=@vif10queue \
1006 queues:234=@vif20queue -- \
1007 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
1008 --id=@vif20queue create queue other-config:max-rate=20000000
1010 At this point, bridge br0 is configured with the ports and eth0 is
1011 configured with the queues that you need for QoS, but nothing is
1012 actually directing packets from vif1.0 or vif2.0 to the queues that
1013 we have set up for them. That means that all of the packets to
1014 eth0 are going to the "default queue", which is not what we want.
1016 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
1017 queues reserved for them:
1019 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
1020 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
1022 Each of the above flows matches on the input port, sets up the
1023 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
1024 executes the "normal" action, which performs the same switching
1025 that Open vSwitch would have done without any OpenFlow flows being
1026 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1027 numbers 5 and 6, respectively, because we set their ofport_request
1028 columns above. If we had not done that, then we would have needed
1029 to find out their port numbers before setting up these flows.)
1031 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1034 By the way, if you delete the bridge created by the above commands,
1037 ovs-vsctl del-br br0
1039 then that will leave one unreferenced QoS record and two
1040 unreferenced Queue records in the Open vSwich database. One way to
1041 clear them out, assuming you don't have other QoS or Queue records
1042 that you want to keep, is:
1044 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1046 If you do want to keep some QoS or Queue records, or the Open
1047 vSwitch you are using is older than version 1.8 (which added the
1048 --all option), then you will have to destroy QoS and Queue records
1051 ### Q: How do I configure ingress policing?
1053 A: A policing policy can be configured on an interface to drop packets
1054 that arrive at a higher rate than the configured value. For example,
1055 the following commands will rate-limit traffic that vif1.0 may
1058 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1059 ovs-vsctl set interface vif1.0 ingress_policing_burst=1000
1061 Traffic policing can interact poorly with some network protocols and
1062 can have surprising results. The "Ingress Policing" section of
1063 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1065 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1066 adding records to the QoS and Queue table, but the results aren't
1069 A: Did you install OpenFlow flows that use your queues? This is the
1070 primary way to tell Open vSwitch which queues you want to use. If
1071 you don't do this, then the default queue will be used, which will
1072 probably not have the effect you want.
1074 Refer to the previous question for an example.
1076 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1077 doesn't yet support. How do I do that?
1079 A: Open vSwitch does not implement QoS itself. Instead, it can
1080 configure some, but not all, of the QoS features built into the
1081 Linux kernel. If you need some QoS feature that OVS cannot
1082 configure itself, then the first step is to figure out whether
1083 Linux QoS supports that feature. If it does, then you can submit a
1084 patch to support Open vSwitch configuration for that feature, or
1085 you can use "tc" directly to configure the feature in Linux. (If
1086 Linux QoS doesn't support the feature you want, then first you have
1087 to add that support to Linux.)
1089 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1090 working as well as I expect.
1092 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1095 - Open vSwitch configures a subset of Linux kernel QoS
1096 features, according to what is in OVSDB. It is possible that
1097 this code has bugs. If you believe that this is so, then you
1098 can configure the Linux traffic control (QoS) stack directly
1099 with the "tc" program. If you get better results that way,
1100 you can send a detailed bug report to bugs@openvswitch.org.
1102 It is certain that Open vSwitch cannot configure every Linux
1103 kernel QoS feature. If you need some feature that OVS cannot
1104 configure, then you can also use "tc" directly (or add that
1107 - The Open vSwitch implementation of OpenFlow allows flows to
1108 be directed to particular queues. This is pretty simple and
1109 unlikely to have serious bugs at this point.
1111 However, most problems with QoS on Linux are not bugs in Open
1112 vSwitch at all. They tend to be either configuration errors
1113 (please see the earlier questions in this section) or issues with
1114 the traffic control (QoS) stack in Linux. The Open vSwitch
1115 developers are not experts on Linux traffic control. We suggest
1116 that, if you believe you are encountering a problem with Linux
1117 traffic control, that you consult the tc manpages (e.g. tc(8),
1118 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1119 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1121 ### Q: Does Open vSwitch support OpenFlow meters?
1123 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1124 OpenFlow meters. There is no implementation of meters in the Open
1125 vSwitch software switch (neither the kernel-based nor userspace
1132 ### Q: What's a VLAN?
1134 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1135 partition a single switch into multiple switches. Suppose, for
1136 example, that you have two groups of machines, group A and group B.
1137 You want the machines in group A to be able to talk to each other,
1138 and you want the machine in group B to be able to talk to each
1139 other, but you don't want the machines in group A to be able to
1140 talk to the machines in group B. You can do this with two
1141 switches, by plugging the machines in group A into one switch and
1142 the machines in group B into the other switch.
1144 If you only have one switch, then you can use VLANs to do the same
1145 thing, by configuring the ports for machines in group A as VLAN
1146 "access ports" for one VLAN and the ports for group B as "access
1147 ports" for a different VLAN. The switch will only forward packets
1148 between ports that are assigned to the same VLAN, so this
1149 effectively subdivides your single switch into two independent
1150 switches, one for each group of machines.
1152 So far we haven't said anything about VLAN headers. With access
1153 ports, like we've described so far, no VLAN header is present in
1154 the Ethernet frame. This means that the machines (or switches)
1155 connected to access ports need not be aware that VLANs are
1156 involved, just like in the case where we use two different physical
1159 Now suppose that you have a whole bunch of switches in your
1160 network, instead of just one, and that some machines in group A are
1161 connected directly to both switches 1 and 2. To allow these
1162 machines to talk to each other, you could add an access port for
1163 group A's VLAN to switch 1 and another to switch 2, and then
1164 connect an Ethernet cable between those ports. That works fine,
1165 but it doesn't scale well as the number of switches and the number
1166 of VLANs increases, because you use up a lot of valuable switch
1167 ports just connecting together your VLANs.
1169 This is where VLAN headers come in. Instead of using one cable and
1170 two ports per VLAN to connect a pair of switches, we configure a
1171 port on each switch as a VLAN "trunk port". Packets sent and
1172 received on a trunk port carry a VLAN header that says what VLAN
1173 the packet belongs to, so that only two ports total are required to
1174 connect the switches, regardless of the number of VLANs in use.
1175 Normally, only switches (either physical or virtual) are connected
1176 to a trunk port, not individual hosts, because individual hosts
1177 don't expect to see a VLAN header in the traffic that they receive.
1179 None of the above discussion says anything about particular VLAN
1180 numbers. This is because VLAN numbers are completely arbitrary.
1181 One must only ensure that a given VLAN is numbered consistently
1182 throughout a network and that different VLANs are given different
1183 numbers. (That said, VLAN 0 is usually synonymous with a packet
1184 that has no VLAN header, and VLAN 4095 is reserved.)
1186 ### Q: VLANs don't work.
1188 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1189 bugs. If you are having problems with VLANs that you suspect to be
1190 driver related, then you have several options:
1192 - Upgrade to Linux 3.3 or later.
1194 - Build and install a fixed version of the particular driver
1195 that is causing trouble, if one is available.
1197 - Use a NIC whose driver does not have VLAN problems.
1199 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
1200 that works around bugs in kernel drivers. To enable VLAN
1201 splinters on interface eth0, use the command:
1203 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1205 For VLAN splinters to be effective, Open vSwitch must know
1206 which VLANs are in use. See the "VLAN splinters" section in
1207 the Interface table in ovs-vswitchd.conf.db(5) for details on
1208 how Open vSwitch infers in-use VLANs.
1210 VLAN splinters increase memory use and reduce performance, so
1211 use them only if needed.
1213 - Apply the "vlan workaround" patch from the XenServer kernel
1214 patch queue, build Open vSwitch against this patched kernel,
1215 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1216 workaround for each interface whose driver is buggy.
1218 (This is a nontrivial exercise, so this option is included
1219 only for completeness.)
1221 It is not always easy to tell whether a Linux kernel driver has
1222 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1223 can help you test. See their manpages for details. Of the two
1224 utilities, ovs-test(8) is newer and more thorough, but
1225 ovs-vlan-test(8) may be easier to use.
1227 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1229 A: Do you have VLANs enabled on the physical switch that OVS is
1230 attached to? Make sure that the port is configured to trunk the
1231 VLAN or VLANs that you are using with OVS.
1233 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1234 and to its destination host, but OVS seems to drop incoming return
1237 A: It's possible that you have the VLAN configured on your physical
1238 switch as the "native" VLAN. In this mode, the switch treats
1239 incoming packets either tagged with the native VLAN or untagged as
1240 part of the native VLAN. It may also send outgoing packets in the
1241 native VLAN without a VLAN tag.
1243 If this is the case, you have two choices:
1245 - Change the physical switch port configuration to tag packets
1246 it forwards to OVS with the native VLAN instead of forwarding
1249 - Change the OVS configuration for the physical port to a
1250 native VLAN mode. For example, the following sets up a
1251 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1253 ovs-vsctl add-br br0
1254 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1256 In this situation, "native-untagged" mode will probably work
1257 equally well. Refer to the documentation for the Port table
1258 in ovs-vswitchd.conf.db(5) for more information.
1260 ### Q: I added a pair of VMs on different VLANs, like this:
1262 ovs-vsctl add-br br0
1263 ovs-vsctl add-port br0 eth0
1264 ovs-vsctl add-port br0 tap0 tag=9
1265 ovs-vsctl add-port br0 tap1 tag=10
1267 but the VMs can't access each other, the external network, or the
1270 A: It is to be expected that the VMs can't access each other. VLANs
1271 are a means to partition a network. When you configured tap0 and
1272 tap1 as access ports for different VLANs, you indicated that they
1273 should be isolated from each other.
1275 As for the external network and the Internet, it seems likely that
1276 the machines you are trying to access are not on VLAN 9 (or 10) and
1277 that the Internet is not available on VLAN 9 (or 10).
1279 ### Q: I added a pair of VMs on the same VLAN, like this:
1281 ovs-vsctl add-br br0
1282 ovs-vsctl add-port br0 eth0
1283 ovs-vsctl add-port br0 tap0 tag=9
1284 ovs-vsctl add-port br0 tap1 tag=9
1286 The VMs can access each other, but not the external network or the
1289 A: It seems likely that the machines you are trying to access in the
1290 external network are not on VLAN 9 and that the Internet is not
1291 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1292 trunk VLAN on the upstream switch port to which eth0 is connected.
1294 ### Q: Can I configure an IP address on a VLAN?
1296 A: Yes. Use an "internal port" configured as an access port. For
1297 example, the following configures IP address 192.168.0.7 on VLAN 9.
1298 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1299 they have an 802.1Q header with VLAN 9. Conversely, traffic
1300 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1303 ovs-vsctl add-br br0
1304 ovs-vsctl add-port br0 eth0
1305 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1306 ifconfig vlan9 192.168.0.7
1308 See also the following question.
1310 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1313 ovs-vsctl add-br br0
1314 ovs-vsctl add-port br0 eth0
1315 ifconfig br0 192.168.0.5
1316 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1317 ifconfig vlan9 192.168.0.9
1319 but other hosts that are only on VLAN 0 can reach the IP address
1320 configured on VLAN 9. What's going on?
1322 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1323 approaches to IP address handling in Internet hosts:
1325 - In the "Strong ES Model", where an ES is a host ("End
1326 System"), an IP address is primarily associated with a
1327 particular interface. The host discards packets that arrive
1328 on interface A if they are destined for an IP address that is
1329 configured on interface B. The host never sends packets from
1330 interface A using a source address configured on interface B.
1332 - In the "Weak ES Model", an IP address is primarily associated
1333 with a host. The host accepts packets that arrive on any
1334 interface if they are destined for any of the host's IP
1335 addresses, even if the address is configured on some
1336 interface other than the one on which it arrived. The host
1337 does not restrict itself to sending packets from an IP
1338 address associated with the originating interface.
1340 Linux uses the weak ES model. That means that when packets
1341 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1342 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1343 address, even though they were not received on vlan9, the network
1346 To simulate the strong ES model on Linux, one may add iptables rule
1347 to filter packets based on source and destination address and
1348 adjust ARP configuration with sysctls.
1350 BSD uses the strong ES model.
1352 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1354 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1355 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1356 implementation of the OpenFlow "normal switching" action. By
1357 default, when Open vSwitch isn't connected to a controller and
1358 nothing has been manually configured in the flow table, all traffic
1359 goes through the "normal switching" action. But, if you set up
1360 OpenFlow flows on your own, through a controller or using ovs-ofctl
1361 or through other means, then you have to implement VLAN handling
1364 You can use "normal switching" as a component of your OpenFlow
1365 actions, e.g. by putting "normal" into the lists of actions on
1366 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1367 controller. In situations where this is not suitable, you can
1368 implement VLAN handling yourself, e.g.:
1370 - If a packet comes in on an access port, and the flow table
1371 needs to send it out on a trunk port, then the flow can add
1372 the appropriate VLAN tag with the "mod_vlan_vid" action.
1374 - If a packet comes in on a trunk port, and the flow table
1375 needs to send it out on an access port, then the flow can
1376 strip the VLAN tag with the "strip_vlan" action.
1378 ### Q: I configured ports on a bridge as access ports with different VLAN
1381 ovs-vsctl add-br br0
1382 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1383 ovs-vsctl add-port br0 eth0
1384 ovs-vsctl add-port br0 tap0 tag=9
1385 ovs-vsctl add-port br0 tap1 tag=10
1387 but the VMs running behind tap0 and tap1 can still communicate,
1388 that is, they are not isolated from each other even though they are
1391 A: Do you have a controller configured on br0 (as the commands above
1392 do)? If so, then this is a variant on the previous question, "My
1393 OpenFlow controller doesn't see the VLANs that I expect," and you
1394 can refer to the answer there for more information.
1396 ### Q: How MAC learning works with VLANs?
1398 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1399 OFPP_NORMAL action. I.e. it logically has separate learning tables
1406 ### Q: What's a VXLAN?
1408 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1409 to solve the scaling challenges of VLAN networks in a multi-tenant
1410 environment. VXLAN is an overlay network which transports an L2 network
1411 over an existing L3 network. For more information on VXLAN, please see
1414 http://tools.ietf.org/html/rfc7348
1416 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1418 A: Open vSwitch currently supports the framing format for packets on the
1419 wire. There is currently no support for the multicast aspects of VXLAN.
1420 To get around the lack of multicast support, it is possible to
1421 pre-provision MAC to IP address mappings either manually or from a
1424 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1427 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1428 is 4789. However, it is possible to configure the destination UDP port
1429 manually on a per-VXLAN tunnel basis. An example of this configuration is
1432 ovs-vsctl add-br br0
1433 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1434 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1435 options:dst_port=8472
1438 Using OpenFlow (Manually or Via Controller)
1439 -------------------------------------------
1441 ### Q: What versions of OpenFlow does Open vSwitch support?
1443 A: The following table lists the versions of OpenFlow supported by
1444 each version of Open vSwitch:
1446 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5
1447 ###============ ===== ===== ===== ===== ===== =====
1448 1.9 and earlier yes --- --- --- --- ---
1449 1.10 yes --- [*] [*] --- ---
1450 1.11 yes --- [*] [*] --- ---
1451 2.0 yes [*] [*] [*] --- ---
1452 2.1 yes [*] [*] [*] --- ---
1453 2.2 yes [*] [*] [*] [%] [*]
1454 2.3 yes yes yes yes [*] [*]
1456 [*] Supported, with one or more missing features.
1457 [%] Experimental, unsafe implementation.
1459 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1460 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1461 1.2, and 1.3 must be enabled manually in ovs-vswitchd. OpenFlow
1462 1.4 and 1.5 are also supported, with missing features, in Open
1463 vSwitch 2.3 and later, but not enabled by default. In any case,
1464 the user may override the default:
1466 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1468 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1470 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1472 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1474 - To enable only OpenFlow 1.0 on bridge br0:
1476 ovs-vsctl set bridge br0 protocols=OpenFlow10
1478 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1479 default. Use the -O option to enable support for later versions of
1480 OpenFlow in ovs-ofctl. For example:
1482 ovs-ofctl -O OpenFlow13 dump-flows br0
1484 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1485 1.4 that could cause crashes. We don't recommend enabling it.)
1487 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1488 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1489 1.5 is solidly implemented, Open vSwitch will enable those version
1492 ### Q: Does Open vSwitch support MPLS?
1494 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1495 these versions can match on MPLS Ethernet types, but they cannot
1496 match, push, or pop MPLS labels, nor can they look past MPLS labels
1497 into the encapsulated packet.
1499 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1500 for MPLS. With the userspace datapath only, these versions can
1501 match, push, or pop a single MPLS label, but they still cannot look
1502 past MPLS labels (even after popping them) into the encapsulated
1503 packet. Kernel datapath support is unchanged from earlier
1506 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1507 label and look past the MPLS label into the encapsulated packet.
1508 Both userspace and kernel datapaths will be supported, but MPLS
1509 processing always happens in userspace either way, so kernel
1510 datapath performance will be disappointing.
1512 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1513 labels and look past the MPLS label into the encapsulated packet.
1514 It will have kernel support for MPLS, yielding improved
1517 ### Q: I'm getting "error type 45250 code 0". What's that?
1519 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1520 vSwitch uses this extension when it must report an error to an
1521 OpenFlow controller but no standard OpenFlow error code is
1524 Open vSwitch logs the errors that it sends to controllers, so the
1525 easiest thing to do is probably to look at the ovs-vswitchd log to
1526 find out what the error was.
1528 If you want to dissect the extended error message yourself, the
1529 format is documented in include/openflow/nicira-ext.h in the Open
1530 vSwitch source distribution. The extended error codes are
1531 documented in lib/ofp-errors.h.
1533 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1534 doesn't actually appear through the OpenFlow connection, even
1535 though I know that it's going through.
1536 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1537 to apply to certain traffic, especially traffic between OVS and
1538 the controller itself.
1540 A: By default, Open vSwitch assumes that OpenFlow controllers are
1541 connected "in-band", that is, that the controllers are actually
1542 part of the network that is being controlled. In in-band mode,
1543 Open vSwitch sets up special "hidden" flows to make sure that
1544 traffic can make it back and forth between OVS and the controllers.
1545 These hidden flows are higher priority than any flows that can be
1546 set up through OpenFlow, and they are not visible through normal
1547 OpenFlow flow table dumps.
1549 Usually, the hidden flows are desirable and helpful, but
1550 occasionally they can cause unexpected behavior. You can view the
1551 full OpenFlow flow table, including hidden flows, on bridge br0
1554 ovs-appctl bridge/dump-flows br0
1556 to help you debug. The hidden flows are those with priorities
1557 greater than 65535 (the maximum priority that can be set with
1560 The DESIGN file at the top level of the Open vSwitch source
1561 distribution describes the in-band model in detail.
1563 If your controllers are not actually in-band (e.g. they are on
1564 localhost via 127.0.0.1, or on a separate network), then you should
1565 configure your controllers in "out-of-band" mode. If you have one
1566 controller on bridge br0, then you can configure out-of-band mode
1569 ovs-vsctl set controller br0 connection-mode=out-of-band
1571 ### Q: I configured all my controllers for out-of-band control mode but
1572 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1574 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1575 set-manager"). By default, Open vSwitch assumes that managers need
1576 in-band rules set up on every bridge. You can disable these rules
1579 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1581 This actually disables in-band control entirely for the bridge, as
1582 if all the bridge's controllers were configured for out-of-band
1585 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1587 A: See answer under "VLANs", above.
1589 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1590 but I got a funny message like this:
1592 ofp_util|INFO|normalization changed ofp_match, details:
1593 ofp_util|INFO| pre: nw_dst=192.168.0.1
1596 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1597 match had disappeared, so that the flow ends up matching every
1600 A: The term "normalization" in the log message means that a flow
1601 cannot match on an L3 field without saying what L3 protocol is in
1602 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1603 so the L3 field match was dropped.
1605 In this case, the L3 protocol could be IP or ARP. A correct
1606 command for each possibility is, respectively:
1608 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1612 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1614 Similarly, a flow cannot match on an L4 field without saying what
1615 L4 protocol is in use. For example, the flow match "tp_src=1234"
1616 is, by itself, meaningless and will be ignored. Instead, to match
1617 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1618 source port 1234, write "udp,tp_src=1234".
1620 ### Q: How can I figure out the OpenFlow port number for a given port?
1622 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1623 respond with an OFPT_FEATURES_REPLY that, among other information,
1624 includes a mapping between OpenFlow port names and numbers. From a
1625 command prompt, "ovs-ofctl show br0" makes such a request and
1626 prints the response for switch br0.
1628 The Interface table in the Open vSwitch database also maps OpenFlow
1629 port names to numbers. To print the OpenFlow port number
1630 associated with interface eth0, run:
1632 ovs-vsctl get Interface eth0 ofport
1634 You can print the entire mapping with:
1636 ovs-vsctl -- --columns=name,ofport list Interface
1638 but the output mixes together interfaces from all bridges in the
1639 database, so it may be confusing if more than one bridge exists.
1641 In the Open vSwitch database, ofport value -1 means that the
1642 interface could not be created due to an error. (The Open vSwitch
1643 log should indicate the reason.) ofport value [] (the empty set)
1644 means that the interface hasn't been created yet. The latter is
1645 normally an intermittent condition (unless ovs-vswitchd is not
1648 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1649 run "ovs-dpctl dump-flows" I don't see them.
1651 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1652 won't display the information that you want. You want to use
1653 "ovs-ofctl dump-flows" instead.
1655 ### Q: It looks like each of the interfaces in my bonded port shows up
1656 as an individual OpenFlow port. Is that right?
1658 A: Yes, Open vSwitch makes individual bond interfaces visible as
1659 OpenFlow ports, rather than the bond as a whole. The interfaces
1660 are treated together as a bond for only a few purposes:
1662 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1663 controller is not configured, this happens implicitly to
1666 - Mirrors configured for output to a bonded port.
1668 It would make a lot of sense for Open vSwitch to present a bond as
1669 a single OpenFlow port. If you want to contribute an
1670 implementation of such a feature, please bring it up on the Open
1671 vSwitch development mailing list at dev@openvswitch.org.
1673 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1674 multiple hosts, and other components. The behavior isn't what I
1677 A: To debug network behavior problems, trace the path of a packet,
1678 hop-by-hop, from its origin in one host to a remote host. If
1679 that's correct, then trace the path of the response packet back to
1682 The open source tool called "plotnetcfg" can help to understand the
1683 relationship between the networking devices on a single host.
1685 Usually a simple ICMP echo request and reply ("ping") packet is
1686 good enough. Start by initiating an ongoing "ping" from the origin
1687 host to a remote host. If you are tracking down a connectivity
1688 problem, the "ping" will not display any successful output, but
1689 packets are still being sent. (In this case the packets being sent
1690 are likely ARP rather than ICMP.)
1692 Tools available for tracing include the following:
1694 - "tcpdump" and "wireshark" for observing hops across network
1695 devices, such as Open vSwitch internal devices and physical
1698 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1699 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1700 These tools allow one to observe the actions being taken on
1701 packets in ongoing flows.
1703 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1704 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1705 documentation, and "Why are there so many different ways to
1706 dump flows?" above for some background.
1708 - "ovs-appctl ofproto/trace" to observe the logic behind how
1709 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1710 documentation. You can out more details about a given flow
1711 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1712 a flow from the output into an "ovs-appctl ofproto/trace"
1715 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1716 observe what goes on at these physical hops.
1718 Starting at the origin of a given packet, observe the packet at
1719 each hop in turn. For example, in one plausible scenario, you
1722 1. "tcpdump" the "eth" interface through which an ARP egresses
1723 a VM, from inside the VM.
1725 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1726 ingresses the host machine.
1728 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1729 the host interface through which the ARP egresses the
1730 physical machine. You may need to use "ovs-dpctl show" to
1731 interpret the port numbers. If the output seems surprising,
1732 you can use "ovs-appctl ofproto/trace" to observe details of
1733 how ovs-vswitchd determined the actions in the "ovs-dpctl
1736 4. "tcpdump" the "eth" interface through which the ARP egresses
1737 the physical machine.
1739 5. "tcpdump" the "eth" interface through which the ARP
1740 ingresses the physical machine, at the remote host that
1743 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1744 remote host that receives the ARP and observe the VM "vif"
1745 or "tap" interface to which the flow is directed. Again,
1746 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1748 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1751 8. "tcpdump" the "eth" interface through which the ARP
1752 ingresses a VM, from inside the VM.
1754 It is likely that during one of these steps you will figure out the
1755 problem. If not, then follow the ARP reply back to the origin, in
1758 ### Q: How do I make a flow drop packets?
1760 A: To drop a packet is to receive it without forwarding it. OpenFlow
1761 explicitly specifies forwarding actions. Thus, a flow with an
1762 empty set of actions does not forward packets anywhere, causing
1763 them to be dropped. You can specify an empty set of actions with
1764 "actions=" on the ovs-ofctl command line. For example:
1766 ovs-ofctl add-flow br0 priority=65535,actions=
1768 would cause every packet entering switch br0 to be dropped.
1770 You can write "drop" explicitly if you like. The effect is the
1771 same. Thus, the following command also causes every packet
1772 entering switch br0 to be dropped:
1774 ovs-ofctl add-flow br0 priority=65535,actions=drop
1776 "drop" is not an action, either in OpenFlow or Open vSwitch.
1777 Rather, it is only a way to say that there are no actions.
1779 ### Q: I added a flow to send packets out the ingress port, like this:
1781 ovs-ofctl add-flow br0 in_port=2,actions=2
1783 but OVS drops the packets instead.
1785 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1786 out its ingress port. The rationale is that dropping these packets
1787 makes it harder to loop the network. Sometimes this behavior can
1788 even be convenient, e.g. it is often the desired behavior in a flow
1789 that forwards a packet to several ports ("floods" the packet).
1791 Sometimes one really needs to send a packet out its ingress port
1792 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1793 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1795 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1797 This also works in some circumstances where the flow doesn't match
1798 on the input port. For example, if you know that your switch has
1799 five ports numbered 2 through 6, then the following will send every
1800 received packet out every port, even its ingress port:
1802 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1806 ovs-ofctl add-flow br0 actions=all,in_port
1808 Sometimes, in complicated flow tables with multiple levels of
1809 "resubmit" actions, a flow needs to output to a particular port
1810 that may or may not be the ingress port. It's difficult to take
1811 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1812 provides, as an OpenFlow extension, the ability to modify the
1813 in_port field. Whatever value is currently in the in_port field is
1814 the port to which outputs will be dropped, as well as the
1815 destination for OFPP_IN_PORT. This means that the following will
1816 reliably output to port 2 or to ports 2 through 6, respectively:
1818 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1819 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1821 If the input port is important, then one may save and restore it on
1824 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1825 load:0->NXM_OF_IN_PORT[],\
1827 pop:NXM_OF_IN_PORT[]
1829 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1830 on port 2. I set up flows to forward only traffic destined to the
1831 other host and drop other traffic, like this:
1833 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1834 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1835 priority=0,actions=drop
1837 But it doesn't work--I don't get any connectivity when I do this.
1840 A: These flows drop the ARP packets that IP hosts use to establish IP
1841 connectivity over Ethernet. To solve the problem, add flows to
1842 allow ARP to pass between the hosts:
1844 priority=5,in_port=1,arp,actions=2
1845 priority=5,in_port=2,arp,actions=1
1847 This issue can manifest other ways, too. The following flows that
1848 match on Ethernet addresses instead of IP addresses will also drop
1849 ARP packets, because ARP requests are broadcast instead of being
1850 directed to a specific host:
1852 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1853 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1854 priority=0,actions=drop
1856 The solution already described above will also work in this case.
1857 It may be better to add flows to allow all multicast and broadcast
1860 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1861 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1863 ### Q: My bridge disconnects from my controller on add-port/del-port.
1865 A: Reconfiguring your bridge can change your bridge's datapath-id because
1866 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1867 In that case, Open vSwitch disconnects from controllers because there's
1868 no graceful way to notify controllers about the change of datapath-id.
1870 To avoid the behaviour, you can configure datapath-id manually.
1872 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1874 ### Q: My controller is getting errors about "buffers". What's going on?
1876 A: When a switch sends a packet to an OpenFlow controller using a
1877 "packet-in" message, it can also keep a copy of that packet in a
1878 "buffer", identified by a 32-bit integer "buffer_id". There are
1879 two advantages to buffering. First, when the controller wants to
1880 tell the switch to do something with the buffered packet (with a
1881 "packet-out" OpenFlow request), it does not need to send another
1882 copy of the packet back across the OpenFlow connection, which
1883 reduces the bandwidth cost of the connection and improves latency.
1884 This enables the second advantage: the switch can optionally send
1885 only the first part of the packet to the controller (assuming that
1886 the switch only needs to look at the first few bytes of the
1887 packet), further reducing bandwidth and improving latency.
1889 However, buffering introduces some issues of its own. First, any
1890 switch has limited resources, so if the controller does not use a
1891 buffered packet, the switch has to decide how long to keep it
1892 buffered. When many packets are sent to a controller and buffered,
1893 Open vSwitch can discard buffered packets that the controller has
1894 not used after as little as 5 seconds. This means that
1895 controllers, if they make use of packet buffering, should use the
1896 buffered packets promptly. (This includes sending a "packet-out"
1897 with no actions if the controller does not want to do anything with
1898 a buffered packet, to clear the packet buffer and effectively
1901 Second, packet buffers are one-time-use, meaning that a controller
1902 cannot use a single packet buffer in two or more "packet-out"
1903 commands. Open vSwitch will respond with an error to the second
1904 and subsequent "packet-out"s in such a case.
1906 Finally, a common error early in controller development is to try
1907 to use buffer_id 0 in a "packet-out" message as if 0 represented
1908 "no buffered packet". This is incorrect usage: the buffer_id with
1909 this meaning is actually 0xffffffff.
1911 ovs-vswitchd(8) describes some details of Open vSwitch packet
1912 buffering that the OpenFlow specification requires implementations
1919 ### Q: How do I implement a new OpenFlow message?
1921 A: Add your new message to "enum ofpraw" and "enum ofptype" in
1922 lib/ofp-msgs.h, following the existing pattern. Then recompile and
1923 fix all of the new warnings, implementing new functionality for the
1924 new message as needed. (If you configure with --enable-Werror, as
1925 described in [INSTALL.md], then it is impossible to miss any warnings.)
1927 If you need to add an OpenFlow vendor extension message for a
1928 vendor that doesn't yet have any extension messages, then you will
1929 also need to edit build-aux/extract-ofp-msgs.
1931 ### Q: How do I add support for a new field or header?
1933 A: Add new members for your field to "struct flow" in lib/flow.h, and
1934 add new enumerations for your new field to "enum mf_field_id" in
1935 lib/meta-flow.h, following the existing pattern. Also, add support
1936 to miniflow_extract() in lib/flow.c for extracting your new field
1937 from a packet into struct miniflow. Then recompile and fix all of
1938 the new warnings, implementing new functionality for the new field
1939 or header as needed. (If you configure with --enable-Werror, as
1940 described in [INSTALL.md], then it is impossible to miss any
1943 If you want kernel datapath support for your new field, you also
1944 need to modify the kernel module for the operating systems you are
1945 interested in. This isn't mandatory, since fields understood only
1946 by userspace work too (with a performance penalty), so it's
1947 reasonable to start development without it. If you implement
1948 kernel module support for Linux, then the Linux kernel "netdev"
1949 mailing list is the place to submit that support first; please read
1950 up on the Linux kernel development process separately. The Windows
1951 datapath kernel module support, on the other hand, is maintained
1952 within the OVS tree, so patches for that can go directly to
1955 ### Q: How do I add support for a new OpenFlow action?
1957 A: Add your new action to "enum ofp_raw_action_type" in
1958 lib/ofp-actions.c, following the existing pattern. Then recompile
1959 and fix all of the new warnings, implementing new functionality for
1960 the new action as needed. (If you configure with --enable-Werror,
1961 as described in [INSTALL.md], then it is impossible to miss any
1964 If you need to add an OpenFlow vendor extension action for a vendor
1965 that doesn't yet have any extension actions, then you will also
1966 need to edit build-aux/extract-ofp-actions.
1972 bugs@openvswitch.org
1973 http://openvswitch.org/
1975 [PORTING.md]:PORTING.md
1976 [WHY-OVS.md]:WHY-OVS.md
1977 [INSTALL.md]:INSTALL.md
1978 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
1979 [INSTALL.DPDK.md]:INSTALL.DPDK.md