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 3.10 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 4.3
160 | 2.6.x | 3.10 to 4.3
162 Open vSwitch userspace should also work with the Linux kernel module
163 built into Linux 3.3 and later.
165 Open vSwitch userspace is not sensitive to the Linux kernel version.
166 It should build against almost any kernel, certainly against 2.6.32
169 ### Q: Are all features available with all datapaths?
171 A: Open vSwitch supports different datapaths on different platforms. Each
172 datapath has a different feature set: the following tables try to summarize
177 * *Linux upstream*: The datapath implemented by the kernel module shipped
178 with Linux upstream. Since features have been gradually
179 introduced into the kernel, the table mentions the first
180 Linux release whose OVS module supports the feature.
182 * *Linux OVS tree*: The datapath implemented by the Linux kernel module
183 distributed with the OVS source tree. Some features of
184 this module rely on functionality not available in older
185 kernels: in this case the minumum Linux version (against
186 which the feature can be compiled) is listed.
188 * *Userspace*: Also known as DPDK, dpif-netdev or dummy datapath. It is the
189 only datapath that works on NetBSD and FreeBSD.
191 * *Hyper-V*: Also known as the Windows datapath.
193 The following table lists the datapath supported features from
194 an Open vSwitch user's perspective.
196 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
197 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
198 Connection tracking | 4.3 | 3.10 | NO | NO |
199 Tunnel - LISP | NO | YES | NO | NO |
200 Tunnel - STT | NO | 3.5 | NO | YES |
201 Tunnel - GRE | 3.11 | YES | YES | YES |
202 Tunnel - VXLAN | 3.12 | YES | YES | YES |
203 Tunnel - Geneve | 3.18 | YES | YES | NO |
204 QoS - Policing | YES | YES | NO | NO |
205 QoS - Shaping | YES | YES | NO | NO |
206 sFlow | YES | YES | YES | NO |
207 IPFIX | 3.10 | YES | YES | NO |
208 Set action | YES | YES | YES | PARTIAL |
209 NIC Bonding | YES | YES | YES | NO |
210 Multiple VTEPs | YES | YES | YES | NO |
213 * Only a limited set of flow fields is modifiable via the set action by the
215 * The Hyper-V datapath only supports one physical NIC per datapath. This is
216 why bonding is not supported.
217 * The Hyper-V datapath can have at most one IP address configured as a
220 The following table lists features that do not *directly* impact an
221 Open vSwitch user, e.g. because their absence can be hidden by the ofproto
222 layer (usually this comes with a performance penalty).
224 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
225 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
226 SCTP flows | 3.12 | YES | YES | YES |
227 MPLS | 3.19 | YES | YES | YES |
228 UFID | 4.0 | YES | YES | NO |
229 Megaflows | 3.12 | YES | YES | NO |
230 Masked set action | 4.0 | YES | YES | NO |
231 Recirculation | 3.19 | YES | YES | YES |
232 TCP flags matching | 3.13 | YES | YES | NO |
233 Validate flow actions | YES | YES | N/A | NO |
234 Multiple datapaths | YES | YES | YES | NO |
235 Tunnel TSO - STT | N/A | YES | NO | YES |
237 ### Q: What DPDK version does each Open vSwitch release work with?
239 A: The following table lists the DPDK version against which the
240 given versions of Open vSwitch will successfully build.
242 | Open vSwitch | DPDK
243 |:------------:|:-----:
250 ### Q: I get an error like this when I configure Open vSwitch:
252 configure: error: Linux kernel in <dir> is version <x>, but
253 version newer than <y> is not supported (please refer to the
258 A: You have the following options:
260 - Use the Linux kernel module supplied with the kernel that you are
261 using. (See also the following FAQ.)
263 - If there is a newer released version of Open vSwitch, consider
264 building that one, because it may support the kernel that you are
265 building against. (To find out, consult the table in the
268 - The Open vSwitch "master" branch may support the kernel that you
269 are using, so consider building the kernel module from "master".
271 All versions of Open vSwitch userspace are compatible with all
272 versions of the Open vSwitch kernel module, so you do not have to
273 use the kernel module from one source along with the userspace
274 programs from the same source.
276 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
278 A: The kernel module in upstream Linux does not include support for
279 LISP. Work is in progress to add support for LISP to the upstream
280 Linux version of the Open vSwitch kernel module. For now, if you
281 need this feature, use the kernel module from the Open vSwitch
282 distribution instead of the upstream Linux kernel module.
284 Certain features require kernel support to function or to have
285 reasonable performance. If the ovs-vswitchd log file indicates that
286 a feature is not supported, consider upgrading to a newer upstream
287 Linux release or using the kernel module paired with the userspace
290 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
292 A: Support for tunnels was added to the upstream Linux kernel module
293 after the rest of Open vSwitch. As a result, some kernels may contain
294 support for Open vSwitch but not tunnels. The minimum kernel version
295 that supports each tunnel protocol is:
297 | Protocol | Linux Kernel
298 |:--------:|:-------------:
302 | LISP | <not upstream>
303 | STT | <not upstream>
305 If you are using a version of the kernel that is older than the one
306 listed above, it is still possible to use that tunnel protocol. However,
307 you must compile and install the kernel module included with the Open
308 vSwitch distribution rather than the one on your machine. If problems
309 persist after doing this, check to make sure that the module that is
310 loaded is the one you expect.
312 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
314 A: Generating outer UDP checksums requires kernel support that was not
315 part of the initial implementation of these protocols. If using the
316 upstream Linux Open vSwitch module, you must use kernel 4.0 or
317 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
318 support UDP checksums.
320 ### Q: What features are not available when using the userspace datapath?
322 A: Tunnel virtual ports are not supported, as described in the
323 previous answer. It is also not possible to use queue-related
324 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
325 may not be transmitted.
327 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
329 In general, the Open vSwitch userspace should be used with the
330 kernel version included in the same release or with the version
331 from upstream Linux. However, when upgrading between two releases
332 of Open vSwitch it is best to migrate userspace first to reduce
333 the possibility of incompatibilities.
335 ### Q: What happened to the bridge compatibility feature?
337 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
338 When it was enabled, Open vSwitch imitated the interface of the
339 Linux kernel "bridge" module. This allowed users to drop Open
340 vSwitch into environments designed to use the Linux kernel bridge
341 module without adapting the environment to use Open vSwitch.
343 Open vSwitch 1.10 and later do not support bridge compatibility.
344 The feature was dropped because version 1.10 adopted a new internal
345 architecture that made bridge compatibility difficult to maintain.
346 Now that many environments use OVS directly, it would be rarely
349 To use bridge compatibility, install OVS 1.9 or earlier, including
350 the accompanying kernel modules (both the main and bridge
351 compatibility modules), following the instructions that come with
352 the release. Be sure to start the ovs-brcompatd daemon.
358 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
360 A: In networking, the terms "bridge" and "switch" are synonyms. Open
361 vSwitch implements an Ethernet switch, which means that it is also
364 ### Q: What's a VLAN?
366 A: See the "VLAN" section below.
372 ### Q: How do I configure a port as an access port?
374 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
375 the following commands configure br0 with eth0 as a trunk port (the
376 default) and tap0 as an access port for VLAN 9:
379 ovs-vsctl add-port br0 eth0
380 ovs-vsctl add-port br0 tap0 tag=9
382 If you want to configure an already added port as an access port,
383 use "ovs-vsctl set", e.g.:
385 ovs-vsctl set port tap0 tag=9
387 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
389 A: The following commands configure br0 with eth0 and tap0 as trunk
390 ports. All traffic coming in or going out on eth0 or tap0 is also
391 mirrored to tap1; any traffic arriving on tap1 is dropped:
394 ovs-vsctl add-port br0 eth0
395 ovs-vsctl add-port br0 tap0
396 ovs-vsctl add-port br0 tap1 \
397 -- --id=@p get port tap1 \
398 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
399 -- set bridge br0 mirrors=@m
401 To later disable mirroring, run:
403 ovs-vsctl clear bridge br0 mirrors
405 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
407 A: Yes. How you configure it depends on what you mean by "promiscuous
410 - Conventionally, "promiscuous mode" is a feature of a network
411 interface card. Ordinarily, a NIC passes to the CPU only the
412 packets actually destined to its host machine. It discards
413 the rest to avoid wasting memory and CPU cycles. When
414 promiscuous mode is enabled, however, it passes every packet
415 to the CPU. On an old-style shared-media or hub-based
416 network, this allows the host to spy on all packets on the
417 network. But in the switched networks that are almost
418 everywhere these days, promiscuous mode doesn't have much
419 effect, because few packets not destined to a host are
420 delivered to the host's NIC.
422 This form of promiscuous mode is configured in the guest OS of
423 the VMs on your bridge, e.g. with "ifconfig".
425 - The VMware vSwitch uses a different definition of "promiscuous
426 mode". When you configure promiscuous mode on a VMware vNIC,
427 the vSwitch sends a copy of every packet received by the
428 vSwitch to that vNIC. That has a much bigger effect than just
429 enabling promiscuous mode in a guest OS. Rather than getting
430 a few stray packets for which the switch does not yet know the
431 correct destination, the vNIC gets every packet. The effect
432 is similar to replacing the vSwitch by a virtual hub.
434 This "promiscuous mode" is what switches normally call "port
435 mirroring" or "SPAN". For information on how to configure
436 SPAN, see "How do I configure a port as a SPAN port, that is,
437 enable mirroring of all traffic to that port?"
439 ### Q: How do I configure a DPDK port as an access port?
441 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
443 If your version is DPDK-enabled it will support the other-config:dpdk-init
444 configuration in the database and will display lines with "EAL:..."
445 during startup when other_config:dpdk-init is set to 'true'.
447 Secondly, when adding a DPDK port, unlike a system port, the
448 type for the interface must be specified. For example;
451 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
453 Finally, it is required that DPDK port names begin with 'dpdk'.
455 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
458 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
460 A: The following commands configure br0 with eth0 as a trunk port and
461 tap0 as an access port for VLAN 10. All traffic coming in or going
462 out on tap0, as well as traffic coming in or going out on eth0 in
463 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
464 VLAN 10, in cases where one is present, is dropped as part of
468 ovs-vsctl add-port br0 eth0
469 ovs-vsctl add-port br0 tap0 tag=10
471 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
473 -- set bridge br0 mirrors=@m
475 To later disable mirroring, run:
477 ovs-vsctl clear bridge br0 mirrors
479 Mirroring to a VLAN can disrupt a network that contains unmanaged
480 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
481 GRE tunnel has fewer caveats than mirroring to a VLAN and should
482 generally be preferred.
484 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
486 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
487 of the specified output-vlan. This loss of information may make
488 the mirrored traffic too hard to interpret.
490 To mirror multiple VLANs, use the commands above, but specify a
491 comma-separated list of VLANs as the value for select-vlan. To
492 mirror every VLAN, use the commands above, but omit select-vlan and
495 When a packet arrives on a VLAN that is used as a mirror output
496 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
497 floods the packet across all the ports for which the mirror output
498 VLAN is configured. (If an OpenFlow controller is in use, then it
499 can override this behavior through the flow table.) If OVS is used
500 as an intermediate switch, rather than an edge switch, this ensures
501 that the RSPAN traffic is distributed through the network.
503 Mirroring to a VLAN can disrupt a network that contains unmanaged
504 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
505 GRE tunnel has fewer caveats than mirroring to a VLAN and should
506 generally be preferred.
508 ### Q: How do I configure mirroring of all traffic to a GRE tunnel?
510 A: The following commands configure br0 with eth0 and tap0 as trunk
511 ports. All traffic coming in or going out on eth0 or tap0 is also
512 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
513 traffic arriving on gre0 is dropped:
516 ovs-vsctl add-port br0 eth0
517 ovs-vsctl add-port br0 tap0
518 ovs-vsctl add-port br0 gre0 \
519 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
520 -- --id=@p get port gre0 \
521 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
522 -- set bridge br0 mirrors=@m
524 To later disable mirroring and destroy the GRE tunnel:
526 ovs-vsctl clear bridge br0 mirrors
527 ovs-vsctl del-port br0 gre0
529 ### Q: Does Open vSwitch support ERSPAN?
531 A: No. ERSPAN is an undocumented proprietary protocol. As an
532 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
535 ### Q: How do I connect two bridges?
537 A: First, why do you want to do this? Two connected bridges are not
538 much different from a single bridge, so you might as well just have
539 a single bridge with all your ports on it.
541 If you still want to connect two bridges, you can use a pair of
542 patch ports. The following example creates bridges br0 and br1,
543 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
544 and br1 with a pair of patch ports.
547 ovs-vsctl add-port br0 eth0
548 ovs-vsctl add-port br0 tap0
550 ovs-vsctl add-port br1 tap1
552 -- add-port br0 patch0 \
553 -- set interface patch0 type=patch options:peer=patch1 \
554 -- add-port br1 patch1 \
555 -- set interface patch1 type=patch options:peer=patch0
557 Bridges connected with patch ports are much like a single bridge.
558 For instance, if the example above also added eth1 to br1, and both
559 eth0 and eth1 happened to be connected to the same next-hop switch,
560 then you could loop your network just as you would if you added
561 eth0 and eth1 to the same bridge (see the "Configuration Problems"
562 section below for more information).
564 If you are using Open vSwitch 1.9 or an earlier version, then you
565 need to be using the kernel module bundled with Open vSwitch rather
566 than the one that is integrated into Linux 3.3 and later, because
567 Open vSwitch 1.9 and earlier versions need kernel support for patch
568 ports. This also means that in Open vSwitch 1.9 and earlier, patch
569 ports will not work with the userspace datapath, only with the
572 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
574 A: Open vSwitch does not support such a configuration.
575 Bridges always have their local ports.
578 Implementation Details
579 ----------------------
581 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
583 A: Open vSwitch uses different kinds of flows for different purposes:
585 - OpenFlow flows are the most important kind of flow. OpenFlow
586 controllers use these flows to define a switch's policy.
587 OpenFlow flows support wildcards, priorities, and multiple
590 When in-band control is in use, Open vSwitch sets up a few
591 "hidden" flows, with priority higher than a controller or the
592 user can configure, that are not visible via OpenFlow. (See
593 the "Controller" section of the FAQ for more information
596 - The Open vSwitch software switch implementation uses a second
597 kind of flow internally. These flows, called "datapath" or
598 "kernel" flows, do not support priorities and comprise only a
599 single table, which makes them suitable for caching. (Like
600 OpenFlow flows, datapath flows do support wildcarding, in Open
601 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
602 also support different actions and number ports differently.
604 Datapath flows are an implementation detail that is subject to
605 change in future versions of Open vSwitch. Even with the
606 current version of Open vSwitch, hardware switch
607 implementations do not necessarily use this architecture.
609 Users and controllers directly control only the OpenFlow flow
610 table. Open vSwitch manages the datapath flow table itself, so
611 users should not normally be concerned with it.
613 ### Q: Why are there so many different ways to dump flows?
615 A: Open vSwitch has two kinds of flows (see the previous question), so
616 it has commands with different purposes for dumping each kind of
619 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
620 hidden flows. This is the most commonly useful form of flow
621 dump. (Unlike the other commands, this should work with any
622 OpenFlow switch, not just Open vSwitch.)
624 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
625 including hidden flows. This is occasionally useful for
626 troubleshooting suspected issues with in-band control.
628 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
629 entries for a Linux kernel-based datapath. In Open vSwitch
630 1.10 and later, ovs-vswitchd merges multiple switches into a
631 single datapath, so it will show all the flows on all your
632 kernel-based switches. This command can occasionally be
633 useful for debugging.
635 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
636 dumps datapath flows for only the specified bridge, regardless
639 ### Q: How does multicast snooping works with VLANs?
641 A: Open vSwitch maintains snooping tables for each VLAN.
643 ### Q: Can OVS populate the kernel flow table in advance instead of in reaction to packets?
645 A: No. There are several reasons:
647 - Kernel flows are not as sophisticated as OpenFlow flows, which
648 means that some OpenFlow policies could require a large number of
649 kernel flows. The "conjunctive match" feature is an extreme
650 example: the number of kernel flows it requires is the product of
651 the number of flows in each dimension.
653 - With multiple OpenFlow flow tables and simple sets of actions, the
654 number of kernel flows required can be as large as the product of
655 the number of flows in each dimension. With more sophisticated
656 actions, the number of kernel flows could be even larger.
658 - Open vSwitch is designed so that any version of OVS userspace
659 interoperates with any version of the OVS kernel module. This
660 forward and backward compatibility requires that userspace observe
661 how the kernel module parses received packets. This is only
662 possible in a straightforward way when userspace adds kernel flows
663 in reaction to received packets.
665 For more relevant information on the architecture of Open vSwitch,
666 please read "The Design and Implementation of Open vSwitch",
667 published in USENIX NSDI 2015.
673 ### Q: I just upgraded and I see a performance drop. Why?
675 A: The OVS kernel datapath may have been updated to a newer version than
676 the OVS userspace components. Sometimes new versions of OVS kernel
677 module add functionality that is backwards compatible with older
678 userspace components but may cause a drop in performance with them.
679 Especially, if a kernel module from OVS 2.1 or newer is paired with
680 OVS userspace 1.10 or older, there will be a performance drop for
683 Updating the OVS userspace components to the latest released
684 version should fix the performance degradation.
686 To get the best possible performance and functionality, it is
687 recommended to pair the same versions of the kernel module and OVS
691 Configuration Problems
692 ----------------------
694 ### Q: I created a bridge and added my Ethernet port to it, using commands
698 ovs-vsctl add-port br0 eth0
700 and as soon as I ran the "add-port" command I lost all connectivity
703 A: A physical Ethernet device that is part of an Open vSwitch bridge
704 should not have an IP address. If one does, then that IP address
705 will not be fully functional.
707 You can restore functionality by moving the IP address to an Open
708 vSwitch "internal" device, such as the network device named after
709 the bridge itself. For example, assuming that eth0's IP address is
710 192.168.128.5, you could run the commands below to fix up the
713 ifconfig eth0 0.0.0.0
714 ifconfig br0 192.168.128.5
716 (If your only connection to the machine running OVS is through the
717 IP address in question, then you would want to run all of these
718 commands on a single command line, or put them into a script.) If
719 there were any additional routes assigned to eth0, then you would
720 also want to use commands to adjust these routes to go through br0.
722 If you use DHCP to obtain an IP address, then you should kill the
723 DHCP client that was listening on the physical Ethernet interface
724 (e.g. eth0) and start one listening on the internal interface
725 (e.g. br0). You might still need to manually clear the IP address
726 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
728 There is no compelling reason why Open vSwitch must work this way.
729 However, this is the way that the Linux kernel bridge module has
730 always worked, so it's a model that those accustomed to Linux
731 bridging are already used to. Also, the model that most people
732 expect is not implementable without kernel changes on all the
733 versions of Linux that Open vSwitch supports.
735 By the way, this issue is not specific to physical Ethernet
736 devices. It applies to all network devices except Open vSwitch
739 ### Q: I created a bridge and added a couple of Ethernet ports to it,
740 ### using commands like these:
743 ovs-vsctl add-port br0 eth0
744 ovs-vsctl add-port br0 eth1
746 and now my network seems to have melted: connectivity is unreliable
747 (even connectivity that doesn't go through Open vSwitch), all the
748 LEDs on my physical switches are blinking, wireshark shows
749 duplicated packets, and CPU usage is very high.
751 A: More than likely, you've looped your network. Probably, eth0 and
752 eth1 are connected to the same physical Ethernet switch. This
753 yields a scenario where OVS receives a broadcast packet on eth0 and
754 sends it out on eth1, then the physical switch connected to eth1
755 sends the packet back on eth0, and so on forever. More complicated
756 scenarios, involving a loop through multiple switches, are possible
759 The solution depends on what you are trying to do:
761 - If you added eth0 and eth1 to get higher bandwidth or higher
762 reliability between OVS and your physical Ethernet switch,
763 use a bond. The following commands create br0 and then add
764 eth0 and eth1 as a bond:
767 ovs-vsctl add-bond br0 bond0 eth0 eth1
769 Bonds have tons of configuration options. Please read the
770 documentation on the Port table in ovs-vswitchd.conf.db(5)
773 Configuration for DPDK-enabled interfaces is slightly less
774 straightforward: see [INSTALL.DPDK.md].
776 - Perhaps you don't actually need eth0 and eth1 to be on the
777 same bridge. For example, if you simply want to be able to
778 connect each of them to virtual machines, then you can put
779 each of them on a bridge of its own:
782 ovs-vsctl add-port br0 eth0
785 ovs-vsctl add-port br1 eth1
787 and then connect VMs to br0 and br1. (A potential
788 disadvantage is that traffic cannot directly pass between br0
789 and br1. Instead, it will go out eth0 and come back in eth1,
792 - If you have a redundant or complex network topology and you
793 want to prevent loops, turn on spanning tree protocol (STP).
794 The following commands create br0, enable STP, and add eth0
795 and eth1 to the bridge. The order is important because you
796 don't want have to have a loop in your network even
800 ovs-vsctl set bridge br0 stp_enable=true
801 ovs-vsctl add-port br0 eth0
802 ovs-vsctl add-port br0 eth1
804 The Open vSwitch implementation of STP is not well tested.
805 Please report any bugs you observe, but if you'd rather avoid
806 acting as a beta tester then another option might be your
809 ### Q: I can't seem to use Open vSwitch in a wireless network.
811 A: Wireless base stations generally only allow packets with the source
812 MAC address of NIC that completed the initial handshake.
813 Therefore, without MAC rewriting, only a single device can
814 communicate over a single wireless link.
816 This isn't specific to Open vSwitch, it's enforced by the access
817 point, so the same problems will show up with the Linux bridge or
818 any other way to do bridging.
820 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
822 A: PPP most commonly carries IP packets, but Open vSwitch works only
823 with Ethernet frames. The correct way to interface PPP to an
824 Ethernet network is usually to use routing instead of switching.
826 ### Q: Is there any documentation on the database tables and fields?
828 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
830 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
831 I only see a datapath called "ovs-system". How can I see datapath
832 information about a particular bridge?
834 A: In version 1.9.0, OVS switched to using a single datapath that is
835 shared by all bridges of that type. The "ovs-appctl dpif/*"
836 commands provide similar functionality that is scoped by the bridge.
838 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
839 see the port in the datapath?
841 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
842 cannot be loaded at the same time. It is likely that on your system the
843 Linux GRE module is already loaded and blocking OVS (to confirm, check
844 dmesg for errors regarding GRE registration). To fix this, unload all
845 GRE modules that appear in lsmod as well as the OVS kernel module. You
846 can then reload the OVS module following the directions in
847 [INSTALL.md], which will ensure that dependencies are satisfied.
849 ### Q: Open vSwitch does not seem to obey my packet filter rules.
851 A: It depends on mechanisms and configurations you want to use.
853 You cannot usefully use typical packet filters, like iptables, on
854 physical Ethernet ports that you add to an Open vSwitch bridge.
855 This is because Open vSwitch captures packets from the interface at
856 a layer lower below where typical packet-filter implementations
857 install their hooks. (This actually applies to any interface of
858 type "system" that you might add to an Open vSwitch bridge.)
860 You can usefully use typical packet filters on Open vSwitch
861 internal ports as they are mostly ordinary interfaces from the point
862 of view of packet filters.
864 For example, suppose you create a bridge br0 and add Ethernet port
865 eth0 to it. Then you can usefully add iptables rules to affect the
866 internal interface br0, but not the physical interface eth0. (br0
867 is also where you would add an IP address, as discussed elsewhere
870 For simple filtering rules, it might be possible to achieve similar
871 results by installing appropriate OpenFlow flows instead.
873 If the use of a particular packet filter setup is essential, Open
874 vSwitch might not be the best choice for you. On Linux, you might
875 want to consider using the Linux Bridge. (This is the only choice if
876 you want to use ebtables rules.) On NetBSD, you might want to
877 consider using the bridge(4) with BRIDGE_IPF option.
879 ### Q: It seems that Open vSwitch does nothing when I removed a port and
880 then immediately put it back. For example, consider that p1 is
881 a port of type=internal:
883 ovs-vsctl del-port br0 p1 -- \
885 set interface p1 type=internal
887 A: It's an expected behaviour.
889 If del-port and add-port happen in a single OVSDB transaction as
890 your example, Open vSwitch always "skips" the intermediate steps.
891 Even if they are done in multiple transactions, it's still allowed
892 for Open vSwitch to skip the intermediate steps and just implement
893 the overall effect. In both cases, your example would be turned
896 If you want to make Open vSwitch actually destroy and then re-create
897 the port for some side effects like resetting kernel setting for the
898 corresponding interface, you need to separate operations into multiple
899 OVSDB transactions and ensure that at least the first one does not have
900 --no-wait. In the following example, the first ovs-vsctl will block
901 until Open vSwitch reloads the new configuration and removes the port:
903 ovs-vsctl del-port br0 p1
904 ovs-vsctl add-port br0 p1 -- \
905 set interface p1 type=internal
907 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
908 it takes too long (minutes or hours) to do it with ovs-vsctl. How
911 A: If you add them one at a time with ovs-vsctl, it can take a long
912 time to add thousands of ports to an Open vSwitch bridge. This is
913 because every invocation of ovs-vsctl first reads the current
914 configuration from OVSDB. As the number of ports grows, this
915 starts to take an appreciable amount of time, and when it is
916 repeated thousands of times the total time becomes significant.
918 The solution is to add the ports in one invocation of ovs-vsctl (or
919 a small number of them). For example, using bash:
922 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
925 takes seconds, not minutes or hours, in the OVS sandbox environment.
927 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
928 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
931 A: Open vSwitch wasn't able to create the bridge. Check the
932 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
933 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
935 In general, the Open vSwitch database reflects the desired
936 configuration state. ovs-vswitchd monitors the database and, when
937 it changes, reconfigures the system to reflect the new desired
938 state. This normally happens very quickly. Thus, a discrepancy
939 between the database and the actual state indicates that
940 ovs-vswitchd could not implement the configuration, and so one
941 should check the log to find out why. (Another possible cause is
942 that ovs-vswitchd is not running. This will make "ovs-vsctl"
943 commands hang, if they change the configuration, unless one
944 specifies "--no-wait".)
946 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
947 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
948 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
951 A: Open vSwitch wasn't able to create the port. Check the
952 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
953 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
954 see the previous question for more information.
956 You may want to upgrade to Open vSwitch 2.3 (or later), in which
957 ovs-vsctl will immediately report when there is an issue creating a
960 ### Q: I created a tap device tap0, configured an IP address on it, and
961 added it to a bridge, like this:
964 ifconfig tap0 192.168.0.123
966 ovs-vsctl add-port br0 tap0
968 I expected that I could then use this IP address to contact other
969 hosts on the network, but it doesn't work. Why not?
971 A: The short answer is that this is a misuse of a "tap" device. Use
972 an "internal" device implemented by Open vSwitch, which works
973 differently and is designed for this use. To solve this problem
974 with an internal device, instead run:
977 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
978 ifconfig int0 192.168.0.123
980 Even more simply, you can take advantage of the internal port that
981 every bridge has under the name of the bridge:
984 ifconfig br0 192.168.0.123
986 In more detail, a "tap" device is an interface between the Linux
987 (or *BSD) network stack and a user program that opens it as a
988 socket. When the "tap" device transmits a packet, it appears in
989 the socket opened by the userspace program. Conversely, when the
990 userspace program writes to the "tap" socket, the kernel TCP/IP
991 stack processes the packet as if it had been received by the "tap"
994 Consider the configuration above. Given this configuration, if you
995 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
996 routing stack will transmit an ARP on the tap0 device. Open
997 vSwitch userspace treats "tap" devices just like any other network
998 device; that is, it doesn't open them as "tap" sockets. That means
999 that the ARP packet will simply get dropped.
1001 You might wonder why the Open vSwitch kernel module doesn't
1002 intercept the ARP packet and bridge it. After all, Open vSwitch
1003 intercepts packets on other devices. The answer is that Open
1004 vSwitch only intercepts *received* packets, but this is a packet
1005 being transmitted. The same thing happens for all other types of
1006 network devices, except for Open vSwitch "internal" ports. If you,
1007 for example, add a physical Ethernet port to an OVS bridge,
1008 configure an IP address on a physical Ethernet port, and then issue
1009 a "ping" to an address in that subnet, the same thing happens: an
1010 ARP gets transmitted on the physical Ethernet port and Open vSwitch
1011 never sees it. (You should not do that, as documented at the
1012 beginning of this section.)
1014 It can make sense to add a "tap" device to an Open vSwitch bridge,
1015 if some userspace program (other than Open vSwitch) has opened the
1016 tap socket. This is the case, for example, if the "tap" device was
1017 created by KVM (or QEMU) to simulate a virtual NIC. In such a
1018 case, when OVS bridges a packet to the "tap" device, the kernel
1019 forwards that packet to KVM in userspace, which passes it along to
1020 the VM, and in the other direction, when the VM sends a packet, KVM
1021 writes it to the "tap" socket, which causes OVS to receive it and
1022 bridge it to the other OVS ports. Please note that in such a case
1023 no IP address is configured on the "tap" device (there is normally
1024 an IP address configured in the virtual NIC inside the VM, but this
1025 is not visible to the host Linux kernel or to Open vSwitch).
1027 There is one special case in which Open vSwitch does directly read
1028 and write "tap" sockets. This is an implementation detail of the
1029 Open vSwitch userspace switch, which implements its "internal"
1030 ports as Linux (or *BSD) "tap" sockets. In such a userspace
1031 switch, OVS receives packets sent on the "tap" device used to
1032 implement an "internal" port by reading the associated "tap"
1033 socket, and bridges them to the rest of the switch. In the other
1034 direction, OVS transmits packets bridged to the "internal" port by
1035 writing them to the "tap" socket, causing them to be processed by
1036 the kernel TCP/IP stack as if they had been received on the "tap"
1037 device. Users should not need to be concerned with this
1038 implementation detail.
1040 Open vSwitch has a network device type called "tap". This is
1041 intended only for implementing "internal" ports in the OVS
1042 userspace switch and should not be used otherwise. In particular,
1043 users should not configure KVM "tap" devices as type "tap" (use
1044 type "system", the default, instead).
1047 Quality of Service (QoS)
1048 ------------------------
1050 ### Q: Does OVS support Quality of Service (QoS)?
1052 A: Yes. For traffic that egresses from a switch, OVS supports traffic
1053 shaping; for traffic that ingresses into a switch, OVS support
1054 policing. Policing is a simple form of quality-of-service that
1055 simply drops packets received in excess of the configured rate. Due
1056 to its simplicity, policing is usually less accurate and less
1057 effective than egress traffic shaping, which queues packets.
1059 Keep in mind that ingress and egress are from the perspective of the
1060 switch. That means that egress shaping limits the rate at which
1061 traffic is allowed to transmit from a physical interface, but the
1062 rate at which traffic will be received on a virtual machine's VIF.
1063 For ingress policing, the behavior is the opposite.
1065 ### Q: How do I configure egress traffic shaping?
1067 A: Suppose that you want to set up bridge br0 connected to physical
1068 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
1069 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
1070 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
1071 could configure the bridge this way:
1075 add-port br0 eth0 -- \
1076 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
1077 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
1078 set port eth0 qos=@newqos -- \
1079 --id=@newqos create qos type=linux-htb \
1080 other-config:max-rate=1000000000 \
1081 queues:123=@vif10queue \
1082 queues:234=@vif20queue -- \
1083 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
1084 --id=@vif20queue create queue other-config:max-rate=20000000
1086 At this point, bridge br0 is configured with the ports and eth0 is
1087 configured with the queues that you need for QoS, but nothing is
1088 actually directing packets from vif1.0 or vif2.0 to the queues that
1089 we have set up for them. That means that all of the packets to
1090 eth0 are going to the "default queue", which is not what we want.
1092 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
1093 queues reserved for them:
1095 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
1096 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
1098 Each of the above flows matches on the input port, sets up the
1099 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
1100 executes the "normal" action, which performs the same switching
1101 that Open vSwitch would have done without any OpenFlow flows being
1102 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1103 numbers 5 and 6, respectively, because we set their ofport_request
1104 columns above. If we had not done that, then we would have needed
1105 to find out their port numbers before setting up these flows.)
1107 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1110 By the way, if you delete the bridge created by the above commands,
1113 ovs-vsctl del-br br0
1115 then that will leave one unreferenced QoS record and two
1116 unreferenced Queue records in the Open vSwich database. One way to
1117 clear them out, assuming you don't have other QoS or Queue records
1118 that you want to keep, is:
1120 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1122 If you do want to keep some QoS or Queue records, or the Open
1123 vSwitch you are using is older than version 1.8 (which added the
1124 --all option), then you will have to destroy QoS and Queue records
1127 ### Q: How do I configure ingress policing?
1129 A: A policing policy can be configured on an interface to drop packets
1130 that arrive at a higher rate than the configured value. For example,
1131 the following commands will rate-limit traffic that vif1.0 may
1134 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1135 ovs-vsctl set interface vif1.0 ingress_policing_burst=8000
1137 Traffic policing can interact poorly with some network protocols and
1138 can have surprising results. The "Ingress Policing" section of
1139 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1141 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1142 adding records to the QoS and Queue table, but the results aren't
1145 A: Did you install OpenFlow flows that use your queues? This is the
1146 primary way to tell Open vSwitch which queues you want to use. If
1147 you don't do this, then the default queue will be used, which will
1148 probably not have the effect you want.
1150 Refer to the previous question for an example.
1152 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1153 doesn't yet support. How do I do that?
1155 A: Open vSwitch does not implement QoS itself. Instead, it can
1156 configure some, but not all, of the QoS features built into the
1157 Linux kernel. If you need some QoS feature that OVS cannot
1158 configure itself, then the first step is to figure out whether
1159 Linux QoS supports that feature. If it does, then you can submit a
1160 patch to support Open vSwitch configuration for that feature, or
1161 you can use "tc" directly to configure the feature in Linux. (If
1162 Linux QoS doesn't support the feature you want, then first you have
1163 to add that support to Linux.)
1165 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1166 working as well as I expect.
1168 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1171 - Open vSwitch configures a subset of Linux kernel QoS
1172 features, according to what is in OVSDB. It is possible that
1173 this code has bugs. If you believe that this is so, then you
1174 can configure the Linux traffic control (QoS) stack directly
1175 with the "tc" program. If you get better results that way,
1176 you can send a detailed bug report to bugs@openvswitch.org.
1178 It is certain that Open vSwitch cannot configure every Linux
1179 kernel QoS feature. If you need some feature that OVS cannot
1180 configure, then you can also use "tc" directly (or add that
1183 - The Open vSwitch implementation of OpenFlow allows flows to
1184 be directed to particular queues. This is pretty simple and
1185 unlikely to have serious bugs at this point.
1187 However, most problems with QoS on Linux are not bugs in Open
1188 vSwitch at all. They tend to be either configuration errors
1189 (please see the earlier questions in this section) or issues with
1190 the traffic control (QoS) stack in Linux. The Open vSwitch
1191 developers are not experts on Linux traffic control. We suggest
1192 that, if you believe you are encountering a problem with Linux
1193 traffic control, that you consult the tc manpages (e.g. tc(8),
1194 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1195 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1197 ### Q: Does Open vSwitch support OpenFlow meters?
1199 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1200 OpenFlow meters. There is no implementation of meters in the Open
1201 vSwitch software switch (neither the kernel-based nor userspace
1208 ### Q: What's a VLAN?
1210 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1211 partition a single switch into multiple switches. Suppose, for
1212 example, that you have two groups of machines, group A and group B.
1213 You want the machines in group A to be able to talk to each other,
1214 and you want the machine in group B to be able to talk to each
1215 other, but you don't want the machines in group A to be able to
1216 talk to the machines in group B. You can do this with two
1217 switches, by plugging the machines in group A into one switch and
1218 the machines in group B into the other switch.
1220 If you only have one switch, then you can use VLANs to do the same
1221 thing, by configuring the ports for machines in group A as VLAN
1222 "access ports" for one VLAN and the ports for group B as "access
1223 ports" for a different VLAN. The switch will only forward packets
1224 between ports that are assigned to the same VLAN, so this
1225 effectively subdivides your single switch into two independent
1226 switches, one for each group of machines.
1228 So far we haven't said anything about VLAN headers. With access
1229 ports, like we've described so far, no VLAN header is present in
1230 the Ethernet frame. This means that the machines (or switches)
1231 connected to access ports need not be aware that VLANs are
1232 involved, just like in the case where we use two different physical
1235 Now suppose that you have a whole bunch of switches in your
1236 network, instead of just one, and that some machines in group A are
1237 connected directly to both switches 1 and 2. To allow these
1238 machines to talk to each other, you could add an access port for
1239 group A's VLAN to switch 1 and another to switch 2, and then
1240 connect an Ethernet cable between those ports. That works fine,
1241 but it doesn't scale well as the number of switches and the number
1242 of VLANs increases, because you use up a lot of valuable switch
1243 ports just connecting together your VLANs.
1245 This is where VLAN headers come in. Instead of using one cable and
1246 two ports per VLAN to connect a pair of switches, we configure a
1247 port on each switch as a VLAN "trunk port". Packets sent and
1248 received on a trunk port carry a VLAN header that says what VLAN
1249 the packet belongs to, so that only two ports total are required to
1250 connect the switches, regardless of the number of VLANs in use.
1251 Normally, only switches (either physical or virtual) are connected
1252 to a trunk port, not individual hosts, because individual hosts
1253 don't expect to see a VLAN header in the traffic that they receive.
1255 None of the above discussion says anything about particular VLAN
1256 numbers. This is because VLAN numbers are completely arbitrary.
1257 One must only ensure that a given VLAN is numbered consistently
1258 throughout a network and that different VLANs are given different
1259 numbers. (That said, VLAN 0 is usually synonymous with a packet
1260 that has no VLAN header, and VLAN 4095 is reserved.)
1262 ### Q: VLANs don't work.
1264 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1265 bugs. If you are having problems with VLANs that you suspect to be
1266 driver related, then you have several options:
1268 - Upgrade to Linux 3.3 or later.
1270 - Build and install a fixed version of the particular driver
1271 that is causing trouble, if one is available.
1273 - Use a NIC whose driver does not have VLAN problems.
1275 - Use "VLAN splinters", a feature in Open vSwitch 1.4 upto 2.5
1276 that works around bugs in kernel drivers. To enable VLAN
1277 splinters on interface eth0, use the command:
1279 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1281 For VLAN splinters to be effective, Open vSwitch must know
1282 which VLANs are in use. See the "VLAN splinters" section in
1283 the Interface table in ovs-vswitchd.conf.db(5) for details on
1284 how Open vSwitch infers in-use VLANs.
1286 VLAN splinters increase memory use and reduce performance, so
1287 use them only if needed.
1289 - Apply the "vlan workaround" patch from the XenServer kernel
1290 patch queue, build Open vSwitch against this patched kernel,
1291 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1292 workaround for each interface whose driver is buggy.
1294 (This is a nontrivial exercise, so this option is included
1295 only for completeness.)
1297 It is not always easy to tell whether a Linux kernel driver has
1298 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1299 can help you test. See their manpages for details. Of the two
1300 utilities, ovs-test(8) is newer and more thorough, but
1301 ovs-vlan-test(8) may be easier to use.
1303 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1305 A: Do you have VLANs enabled on the physical switch that OVS is
1306 attached to? Make sure that the port is configured to trunk the
1307 VLAN or VLANs that you are using with OVS.
1309 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1310 and to its destination host, but OVS seems to drop incoming return
1313 A: It's possible that you have the VLAN configured on your physical
1314 switch as the "native" VLAN. In this mode, the switch treats
1315 incoming packets either tagged with the native VLAN or untagged as
1316 part of the native VLAN. It may also send outgoing packets in the
1317 native VLAN without a VLAN tag.
1319 If this is the case, you have two choices:
1321 - Change the physical switch port configuration to tag packets
1322 it forwards to OVS with the native VLAN instead of forwarding
1325 - Change the OVS configuration for the physical port to a
1326 native VLAN mode. For example, the following sets up a
1327 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1329 ovs-vsctl add-br br0
1330 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1332 In this situation, "native-untagged" mode will probably work
1333 equally well. Refer to the documentation for the Port table
1334 in ovs-vswitchd.conf.db(5) for more information.
1336 ### Q: I added a pair of VMs on different VLANs, like this:
1338 ovs-vsctl add-br br0
1339 ovs-vsctl add-port br0 eth0
1340 ovs-vsctl add-port br0 tap0 tag=9
1341 ovs-vsctl add-port br0 tap1 tag=10
1343 but the VMs can't access each other, the external network, or the
1346 A: It is to be expected that the VMs can't access each other. VLANs
1347 are a means to partition a network. When you configured tap0 and
1348 tap1 as access ports for different VLANs, you indicated that they
1349 should be isolated from each other.
1351 As for the external network and the Internet, it seems likely that
1352 the machines you are trying to access are not on VLAN 9 (or 10) and
1353 that the Internet is not available on VLAN 9 (or 10).
1355 ### Q: I added a pair of VMs on the same VLAN, like this:
1357 ovs-vsctl add-br br0
1358 ovs-vsctl add-port br0 eth0
1359 ovs-vsctl add-port br0 tap0 tag=9
1360 ovs-vsctl add-port br0 tap1 tag=9
1362 The VMs can access each other, but not the external network or the
1365 A: It seems likely that the machines you are trying to access in the
1366 external network are not on VLAN 9 and that the Internet is not
1367 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1368 trunk VLAN on the upstream switch port to which eth0 is connected.
1370 ### Q: Can I configure an IP address on a VLAN?
1372 A: Yes. Use an "internal port" configured as an access port. For
1373 example, the following configures IP address 192.168.0.7 on VLAN 9.
1374 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1375 they have an 802.1Q header with VLAN 9. Conversely, traffic
1376 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1379 ovs-vsctl add-br br0
1380 ovs-vsctl add-port br0 eth0
1381 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1382 ifconfig vlan9 192.168.0.7
1384 See also the following question.
1386 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1389 ovs-vsctl add-br br0
1390 ovs-vsctl add-port br0 eth0
1391 ifconfig br0 192.168.0.5
1392 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1393 ifconfig vlan9 192.168.0.9
1395 but other hosts that are only on VLAN 0 can reach the IP address
1396 configured on VLAN 9. What's going on?
1398 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1399 approaches to IP address handling in Internet hosts:
1401 - In the "Strong ES Model", where an ES is a host ("End
1402 System"), an IP address is primarily associated with a
1403 particular interface. The host discards packets that arrive
1404 on interface A if they are destined for an IP address that is
1405 configured on interface B. The host never sends packets from
1406 interface A using a source address configured on interface B.
1408 - In the "Weak ES Model", an IP address is primarily associated
1409 with a host. The host accepts packets that arrive on any
1410 interface if they are destined for any of the host's IP
1411 addresses, even if the address is configured on some
1412 interface other than the one on which it arrived. The host
1413 does not restrict itself to sending packets from an IP
1414 address associated with the originating interface.
1416 Linux uses the weak ES model. That means that when packets
1417 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1418 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1419 address, even though they were not received on vlan9, the network
1422 To simulate the strong ES model on Linux, one may add iptables rule
1423 to filter packets based on source and destination address and
1424 adjust ARP configuration with sysctls.
1426 BSD uses the strong ES model.
1428 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1430 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1431 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1432 implementation of the OpenFlow "normal switching" action. By
1433 default, when Open vSwitch isn't connected to a controller and
1434 nothing has been manually configured in the flow table, all traffic
1435 goes through the "normal switching" action. But, if you set up
1436 OpenFlow flows on your own, through a controller or using ovs-ofctl
1437 or through other means, then you have to implement VLAN handling
1440 You can use "normal switching" as a component of your OpenFlow
1441 actions, e.g. by putting "normal" into the lists of actions on
1442 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1443 controller. In situations where this is not suitable, you can
1444 implement VLAN handling yourself, e.g.:
1446 - If a packet comes in on an access port, and the flow table
1447 needs to send it out on a trunk port, then the flow can add
1448 the appropriate VLAN tag with the "mod_vlan_vid" action.
1450 - If a packet comes in on a trunk port, and the flow table
1451 needs to send it out on an access port, then the flow can
1452 strip the VLAN tag with the "strip_vlan" action.
1454 ### Q: I configured ports on a bridge as access ports with different VLAN
1457 ovs-vsctl add-br br0
1458 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1459 ovs-vsctl add-port br0 eth0
1460 ovs-vsctl add-port br0 tap0 tag=9
1461 ovs-vsctl add-port br0 tap1 tag=10
1463 but the VMs running behind tap0 and tap1 can still communicate,
1464 that is, they are not isolated from each other even though they are
1467 A: Do you have a controller configured on br0 (as the commands above
1468 do)? If so, then this is a variant on the previous question, "My
1469 OpenFlow controller doesn't see the VLANs that I expect," and you
1470 can refer to the answer there for more information.
1472 ### Q: How MAC learning works with VLANs?
1474 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1475 OFPP_NORMAL action. I.e. it logically has separate learning tables
1482 ### Q: What's a VXLAN?
1484 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1485 to solve the scaling challenges of VLAN networks in a multi-tenant
1486 environment. VXLAN is an overlay network which transports an L2 network
1487 over an existing L3 network. For more information on VXLAN, please see
1490 http://tools.ietf.org/html/rfc7348
1492 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1494 A: Open vSwitch currently supports the framing format for packets on the
1495 wire. There is currently no support for the multicast aspects of VXLAN.
1496 To get around the lack of multicast support, it is possible to
1497 pre-provision MAC to IP address mappings either manually or from a
1500 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1503 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1504 is 4789. However, it is possible to configure the destination UDP port
1505 manually on a per-VXLAN tunnel basis. An example of this configuration is
1508 ovs-vsctl add-br br0
1509 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1510 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1511 options:dst_port=8472
1514 Using OpenFlow (Manually or Via Controller)
1515 -------------------------------------------
1517 ### Q: What versions of OpenFlow does Open vSwitch support?
1519 A: The following table lists the versions of OpenFlow supported by
1520 each version of Open vSwitch:
1522 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5 OF1.6
1523 ###============ ===== ===== ===== ===== ===== ===== =====
1524 1.9 and earlier yes --- --- --- --- --- ---
1525 1.10 yes --- [*] [*] --- --- ---
1526 1.11 yes --- [*] [*] --- --- ---
1527 2.0 yes [*] [*] [*] --- --- ---
1528 2.1 yes [*] [*] [*] --- --- ---
1529 2.2 yes [*] [*] [*] [%] [*] ---
1530 2.3 yes yes yes yes [*] [*] ---
1531 2.4 yes yes yes yes [*] [*] ---
1532 2.5 yes yes yes yes [*] [*] [*]
1534 [*] Supported, with one or more missing features.
1535 [%] Experimental, unsafe implementation.
1537 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1538 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1539 1.2, and 1.3 must be enabled manually in ovs-vswitchd.
1541 Some versions of OpenFlow are supported with missing features and
1542 therefore not enabled by default: OpenFlow 1.4 and 1.5, in Open
1543 vSwitch 2.3 and later, as well as OpenFlow 1.6 in Open vSwitch 2.5
1544 and later. Also, the OpenFlow 1.6 specification is still under
1545 development and thus subject to change.
1547 In any case, the user may override the default:
1549 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1551 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1553 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1555 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1557 - To enable only OpenFlow 1.0 on bridge br0:
1559 ovs-vsctl set bridge br0 protocols=OpenFlow10
1561 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1562 default. Use the -O option to enable support for later versions of
1563 OpenFlow in ovs-ofctl. For example:
1565 ovs-ofctl -O OpenFlow13 dump-flows br0
1567 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1568 1.4 that could cause crashes. We don't recommend enabling it.)
1570 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1571 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1572 1.5 is solidly implemented, Open vSwitch will enable those version
1575 ### Q: Does Open vSwitch support MPLS?
1577 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1578 these versions can match on MPLS Ethernet types, but they cannot
1579 match, push, or pop MPLS labels, nor can they look past MPLS labels
1580 into the encapsulated packet.
1582 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1583 for MPLS. With the userspace datapath only, these versions can
1584 match, push, or pop a single MPLS label, but they still cannot look
1585 past MPLS labels (even after popping them) into the encapsulated
1586 packet. Kernel datapath support is unchanged from earlier
1589 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1590 label and look past the MPLS label into the encapsulated packet.
1591 Both userspace and kernel datapaths will be supported, but MPLS
1592 processing always happens in userspace either way, so kernel
1593 datapath performance will be disappointing.
1595 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1596 labels and look past the MPLS label into the encapsulated packet.
1597 It will have kernel support for MPLS, yielding improved
1600 ### Q: I'm getting "error type 45250 code 0". What's that?
1602 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1603 vSwitch uses this extension when it must report an error to an
1604 OpenFlow controller but no standard OpenFlow error code is
1607 Open vSwitch logs the errors that it sends to controllers, so the
1608 easiest thing to do is probably to look at the ovs-vswitchd log to
1609 find out what the error was.
1611 If you want to dissect the extended error message yourself, the
1612 format is documented in include/openflow/nicira-ext.h in the Open
1613 vSwitch source distribution. The extended error codes are
1614 documented in include/openvswitch/ofp-errors.h.
1616 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1617 doesn't actually appear through the OpenFlow connection, even
1618 though I know that it's going through.
1619 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1620 to apply to certain traffic, especially traffic between OVS and
1621 the controller itself.
1623 A: By default, Open vSwitch assumes that OpenFlow controllers are
1624 connected "in-band", that is, that the controllers are actually
1625 part of the network that is being controlled. In in-band mode,
1626 Open vSwitch sets up special "hidden" flows to make sure that
1627 traffic can make it back and forth between OVS and the controllers.
1628 These hidden flows are higher priority than any flows that can be
1629 set up through OpenFlow, and they are not visible through normal
1630 OpenFlow flow table dumps.
1632 Usually, the hidden flows are desirable and helpful, but
1633 occasionally they can cause unexpected behavior. You can view the
1634 full OpenFlow flow table, including hidden flows, on bridge br0
1637 ovs-appctl bridge/dump-flows br0
1639 to help you debug. The hidden flows are those with priorities
1640 greater than 65535 (the maximum priority that can be set with
1643 The DESIGN file at the top level of the Open vSwitch source
1644 distribution describes the in-band model in detail.
1646 If your controllers are not actually in-band (e.g. they are on
1647 localhost via 127.0.0.1, or on a separate network), then you should
1648 configure your controllers in "out-of-band" mode. If you have one
1649 controller on bridge br0, then you can configure out-of-band mode
1652 ovs-vsctl set controller br0 connection-mode=out-of-band
1654 ### Q: I configured all my controllers for out-of-band control mode but
1655 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1657 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1658 set-manager"). By default, Open vSwitch assumes that managers need
1659 in-band rules set up on every bridge. You can disable these rules
1662 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1664 This actually disables in-band control entirely for the bridge, as
1665 if all the bridge's controllers were configured for out-of-band
1668 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1670 A: See answer under "VLANs", above.
1672 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1673 but I got a funny message like this:
1675 ofp_util|INFO|normalization changed ofp_match, details:
1676 ofp_util|INFO| pre: nw_dst=192.168.0.1
1679 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1680 match had disappeared, so that the flow ends up matching every
1683 A: The term "normalization" in the log message means that a flow
1684 cannot match on an L3 field without saying what L3 protocol is in
1685 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1686 so the L3 field match was dropped.
1688 In this case, the L3 protocol could be IP or ARP. A correct
1689 command for each possibility is, respectively:
1691 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1695 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1697 Similarly, a flow cannot match on an L4 field without saying what
1698 L4 protocol is in use. For example, the flow match "tp_src=1234"
1699 is, by itself, meaningless and will be ignored. Instead, to match
1700 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1701 source port 1234, write "udp,tp_src=1234".
1703 ### Q: How can I figure out the OpenFlow port number for a given port?
1705 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1706 respond with an OFPT_FEATURES_REPLY that, among other information,
1707 includes a mapping between OpenFlow port names and numbers. From a
1708 command prompt, "ovs-ofctl show br0" makes such a request and
1709 prints the response for switch br0.
1711 The Interface table in the Open vSwitch database also maps OpenFlow
1712 port names to numbers. To print the OpenFlow port number
1713 associated with interface eth0, run:
1715 ovs-vsctl get Interface eth0 ofport
1717 You can print the entire mapping with:
1719 ovs-vsctl -- --columns=name,ofport list Interface
1721 but the output mixes together interfaces from all bridges in the
1722 database, so it may be confusing if more than one bridge exists.
1724 In the Open vSwitch database, ofport value -1 means that the
1725 interface could not be created due to an error. (The Open vSwitch
1726 log should indicate the reason.) ofport value [] (the empty set)
1727 means that the interface hasn't been created yet. The latter is
1728 normally an intermittent condition (unless ovs-vswitchd is not
1731 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1732 run "ovs-dpctl dump-flows" I don't see them.
1734 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1735 won't display the information that you want. You want to use
1736 "ovs-ofctl dump-flows" instead.
1738 ### Q: It looks like each of the interfaces in my bonded port shows up
1739 as an individual OpenFlow port. Is that right?
1741 A: Yes, Open vSwitch makes individual bond interfaces visible as
1742 OpenFlow ports, rather than the bond as a whole. The interfaces
1743 are treated together as a bond for only a few purposes:
1745 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1746 controller is not configured, this happens implicitly to
1749 - Mirrors configured for output to a bonded port.
1751 It would make a lot of sense for Open vSwitch to present a bond as
1752 a single OpenFlow port. If you want to contribute an
1753 implementation of such a feature, please bring it up on the Open
1754 vSwitch development mailing list at dev@openvswitch.org.
1756 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1757 multiple hosts, and other components. The behavior isn't what I
1760 A: To debug network behavior problems, trace the path of a packet,
1761 hop-by-hop, from its origin in one host to a remote host. If
1762 that's correct, then trace the path of the response packet back to
1765 The open source tool called "plotnetcfg" can help to understand the
1766 relationship between the networking devices on a single host.
1768 Usually a simple ICMP echo request and reply ("ping") packet is
1769 good enough. Start by initiating an ongoing "ping" from the origin
1770 host to a remote host. If you are tracking down a connectivity
1771 problem, the "ping" will not display any successful output, but
1772 packets are still being sent. (In this case the packets being sent
1773 are likely ARP rather than ICMP.)
1775 Tools available for tracing include the following:
1777 - "tcpdump" and "wireshark" for observing hops across network
1778 devices, such as Open vSwitch internal devices and physical
1781 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1782 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1783 These tools allow one to observe the actions being taken on
1784 packets in ongoing flows.
1786 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1787 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1788 documentation, and "Why are there so many different ways to
1789 dump flows?" above for some background.
1791 - "ovs-appctl ofproto/trace" to observe the logic behind how
1792 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1793 documentation. You can out more details about a given flow
1794 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1795 a flow from the output into an "ovs-appctl ofproto/trace"
1798 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1799 observe what goes on at these physical hops.
1801 Starting at the origin of a given packet, observe the packet at
1802 each hop in turn. For example, in one plausible scenario, you
1805 1. "tcpdump" the "eth" interface through which an ARP egresses
1806 a VM, from inside the VM.
1808 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1809 ingresses the host machine.
1811 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1812 the host interface through which the ARP egresses the
1813 physical machine. You may need to use "ovs-dpctl show" to
1814 interpret the port numbers. If the output seems surprising,
1815 you can use "ovs-appctl ofproto/trace" to observe details of
1816 how ovs-vswitchd determined the actions in the "ovs-dpctl
1819 4. "tcpdump" the "eth" interface through which the ARP egresses
1820 the physical machine.
1822 5. "tcpdump" the "eth" interface through which the ARP
1823 ingresses the physical machine, at the remote host that
1826 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1827 remote host that receives the ARP and observe the VM "vif"
1828 or "tap" interface to which the flow is directed. Again,
1829 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1831 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1834 8. "tcpdump" the "eth" interface through which the ARP
1835 ingresses a VM, from inside the VM.
1837 It is likely that during one of these steps you will figure out the
1838 problem. If not, then follow the ARP reply back to the origin, in
1841 ### Q: How do I make a flow drop packets?
1843 A: To drop a packet is to receive it without forwarding it. OpenFlow
1844 explicitly specifies forwarding actions. Thus, a flow with an
1845 empty set of actions does not forward packets anywhere, causing
1846 them to be dropped. You can specify an empty set of actions with
1847 "actions=" on the ovs-ofctl command line. For example:
1849 ovs-ofctl add-flow br0 priority=65535,actions=
1851 would cause every packet entering switch br0 to be dropped.
1853 You can write "drop" explicitly if you like. The effect is the
1854 same. Thus, the following command also causes every packet
1855 entering switch br0 to be dropped:
1857 ovs-ofctl add-flow br0 priority=65535,actions=drop
1859 "drop" is not an action, either in OpenFlow or Open vSwitch.
1860 Rather, it is only a way to say that there are no actions.
1862 ### Q: I added a flow to send packets out the ingress port, like this:
1864 ovs-ofctl add-flow br0 in_port=2,actions=2
1866 but OVS drops the packets instead.
1868 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1869 out its ingress port. The rationale is that dropping these packets
1870 makes it harder to loop the network. Sometimes this behavior can
1871 even be convenient, e.g. it is often the desired behavior in a flow
1872 that forwards a packet to several ports ("floods" the packet).
1874 Sometimes one really needs to send a packet out its ingress port
1875 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1876 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1878 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1880 This also works in some circumstances where the flow doesn't match
1881 on the input port. For example, if you know that your switch has
1882 five ports numbered 2 through 6, then the following will send every
1883 received packet out every port, even its ingress port:
1885 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1889 ovs-ofctl add-flow br0 actions=all,in_port
1891 Sometimes, in complicated flow tables with multiple levels of
1892 "resubmit" actions, a flow needs to output to a particular port
1893 that may or may not be the ingress port. It's difficult to take
1894 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1895 provides, as an OpenFlow extension, the ability to modify the
1896 in_port field. Whatever value is currently in the in_port field is
1897 the port to which outputs will be dropped, as well as the
1898 destination for OFPP_IN_PORT. This means that the following will
1899 reliably output to port 2 or to ports 2 through 6, respectively:
1901 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1902 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1904 If the input port is important, then one may save and restore it on
1907 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1908 load:0->NXM_OF_IN_PORT[],\
1910 pop:NXM_OF_IN_PORT[]
1912 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1913 on port 2. I set up flows to forward only traffic destined to the
1914 other host and drop other traffic, like this:
1916 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1917 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1918 priority=0,actions=drop
1920 But it doesn't work--I don't get any connectivity when I do this.
1923 A: These flows drop the ARP packets that IP hosts use to establish IP
1924 connectivity over Ethernet. To solve the problem, add flows to
1925 allow ARP to pass between the hosts:
1927 priority=5,in_port=1,arp,actions=2
1928 priority=5,in_port=2,arp,actions=1
1930 This issue can manifest other ways, too. The following flows that
1931 match on Ethernet addresses instead of IP addresses will also drop
1932 ARP packets, because ARP requests are broadcast instead of being
1933 directed to a specific host:
1935 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1936 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1937 priority=0,actions=drop
1939 The solution already described above will also work in this case.
1940 It may be better to add flows to allow all multicast and broadcast
1943 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1944 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1946 ### Q: My bridge disconnects from my controller on add-port/del-port.
1948 A: Reconfiguring your bridge can change your bridge's datapath-id because
1949 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1950 In that case, Open vSwitch disconnects from controllers because there's
1951 no graceful way to notify controllers about the change of datapath-id.
1953 To avoid the behaviour, you can configure datapath-id manually.
1955 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1957 ### Q: My controller is getting errors about "buffers". What's going on?
1959 A: When a switch sends a packet to an OpenFlow controller using a
1960 "packet-in" message, it can also keep a copy of that packet in a
1961 "buffer", identified by a 32-bit integer "buffer_id". There are
1962 two advantages to buffering. First, when the controller wants to
1963 tell the switch to do something with the buffered packet (with a
1964 "packet-out" OpenFlow request), it does not need to send another
1965 copy of the packet back across the OpenFlow connection, which
1966 reduces the bandwidth cost of the connection and improves latency.
1967 This enables the second advantage: the switch can optionally send
1968 only the first part of the packet to the controller (assuming that
1969 the switch only needs to look at the first few bytes of the
1970 packet), further reducing bandwidth and improving latency.
1972 However, buffering introduces some issues of its own. First, any
1973 switch has limited resources, so if the controller does not use a
1974 buffered packet, the switch has to decide how long to keep it
1975 buffered. When many packets are sent to a controller and buffered,
1976 Open vSwitch can discard buffered packets that the controller has
1977 not used after as little as 5 seconds. This means that
1978 controllers, if they make use of packet buffering, should use the
1979 buffered packets promptly. (This includes sending a "packet-out"
1980 with no actions if the controller does not want to do anything with
1981 a buffered packet, to clear the packet buffer and effectively
1984 Second, packet buffers are one-time-use, meaning that a controller
1985 cannot use a single packet buffer in two or more "packet-out"
1986 commands. Open vSwitch will respond with an error to the second
1987 and subsequent "packet-out"s in such a case.
1989 Finally, a common error early in controller development is to try
1990 to use buffer_id 0 in a "packet-out" message as if 0 represented
1991 "no buffered packet". This is incorrect usage: the buffer_id with
1992 this meaning is actually 0xffffffff.
1994 ovs-vswitchd(8) describes some details of Open vSwitch packet
1995 buffering that the OpenFlow specification requires implementations
1998 ### Q: How does OVS divide flows among buckets in an OpenFlow "select" group?
2000 A: In Open vSwitch 2.3 and earlier, Open vSwitch used the destination
2001 Ethernet address to choose a bucket in a select group.
2003 Open vSwitch 2.4 and later by default hashes the source and
2004 destination Ethernet address, VLAN ID, Ethernet type, IPv4/v6
2005 source and destination address and protocol, and for TCP and SCTP
2006 only, the source and destination ports. The hash is "symmetric",
2007 meaning that exchanging source and destination addresses does not
2008 change the bucket selection.
2010 Select groups in Open vSwitch 2.4 and later can be configured to
2011 use a different hash function, using a Netronome extension to the
2012 OpenFlow 1.5+ group_mod message. For more information, see
2013 Documentation/group-selection-method-property.txt in the Open
2014 vSwitch source tree. (OpenFlow 1.5 support in Open vSwitch is still
2021 ### Q: How do I implement a new OpenFlow message?
2023 A: Add your new message to "enum ofpraw" and "enum ofptype" in
2024 lib/ofp-msgs.h, following the existing pattern. Then recompile and
2025 fix all of the new warnings, implementing new functionality for the
2026 new message as needed. (If you configure with --enable-Werror, as
2027 described in [INSTALL.md], then it is impossible to miss any warnings.)
2029 If you need to add an OpenFlow vendor extension message for a
2030 vendor that doesn't yet have any extension messages, then you will
2031 also need to edit build-aux/extract-ofp-msgs.
2033 ### Q: How do I add support for a new field or header?
2035 A: Add new members for your field to "struct flow" in lib/flow.h, and
2036 add new enumerations for your new field to "enum mf_field_id" in
2037 lib/meta-flow.h, following the existing pattern. Also, add support
2038 to miniflow_extract() in lib/flow.c for extracting your new field
2039 from a packet into struct miniflow, and to nx_put_raw() in
2040 lib/nx-match.c to output your new field in OXM matches. Then
2041 recompile and fix all of the new warnings, implementing new
2042 functionality for the new field or header as needed. (If you
2043 configure with --enable-Werror, as described in [INSTALL.md], then
2044 it is impossible to miss any warnings.)
2046 If you want kernel datapath support for your new field, you also
2047 need to modify the kernel module for the operating systems you are
2048 interested in. This isn't mandatory, since fields understood only
2049 by userspace work too (with a performance penalty), so it's
2050 reasonable to start development without it. If you implement
2051 kernel module support for Linux, then the Linux kernel "netdev"
2052 mailing list is the place to submit that support first; please read
2053 up on the Linux kernel development process separately. The Windows
2054 datapath kernel module support, on the other hand, is maintained
2055 within the OVS tree, so patches for that can go directly to
2058 ### Q: How do I add support for a new OpenFlow action?
2060 A: Add your new action to "enum ofp_raw_action_type" in
2061 lib/ofp-actions.c, following the existing pattern. Then recompile
2062 and fix all of the new warnings, implementing new functionality for
2063 the new action as needed. (If you configure with --enable-Werror,
2064 as described in [INSTALL.md], then it is impossible to miss any
2067 If you need to add an OpenFlow vendor extension action for a vendor
2068 that doesn't yet have any extension actions, then you will also
2069 need to edit build-aux/extract-ofp-actions.
2075 bugs@openvswitch.org
2076 http://openvswitch.org/
2078 [PORTING.md]:PORTING.md
2079 [WHY-OVS.md]:WHY-OVS.md
2080 [INSTALL.md]:INSTALL.md
2081 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
2082 [INSTALL.DPDK.md]:INSTALL.DPDK.md