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 4.3
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: Are all features available with all versions of the kernel module?
170 A: Features are gradually introduced into the upstream kernel so unless
171 you are running the latest Linux release there may be some missing
172 features. ovs-vswitchd probes for kernel features during startup and
173 lists them in the ovs-vswitchd log file. If you need features that
174 are unsupported, you can upgrade the Linux version or use the kernel
175 module paired with the userspace distribution.
177 For a given Open vSwitch release, the vast majority of features are
178 supported when using the kernel module paired with the userspace
179 distribution, no matter which kernel you build it against. However,
180 from time to time, new features may be introduced which rely on
181 functionality not available in older kernels. The features below are
182 included in the kernel module distributed with OVS, but have more
183 stringent requirements for minimum kernel versions than the support
186 | Feature | Open vSwitch | Linux kernel
187 |:---------:|:------------:|:------------:
189 | Conntrack | 2.5+ | 3.10+
191 ### Q: I get an error like this when I configure Open vSwitch:
193 configure: error: Linux kernel in <dir> is version <x>, but
194 version newer than <y> is not supported (please refer to the
199 A: You have the following options:
201 - Use the Linux kernel module supplied with the kernel that you are
202 using. (See also the following FAQ.)
204 - If there is a newer released version of Open vSwitch, consider
205 building that one, because it may support the kernel that you are
206 building against. (To find out, consult the table in the
209 - The Open vSwitch "master" branch may support the kernel that you
210 are using, so consider building the kernel module from "master".
212 All versions of Open vSwitch userspace are compatible with all
213 versions of the Open vSwitch kernel module, so you do not have to
214 use the kernel module from one source along with the userspace
215 programs from the same source.
217 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
219 A: The kernel module in upstream Linux does not include support for
220 LISP. Work is in progress to add support for LISP to the upstream
221 Linux version of the Open vSwitch kernel module. For now, if you
222 need this feature, use the kernel module from the Open vSwitch
223 distribution instead of the upstream Linux kernel module.
225 Certain features require kernel support to function or to have
226 reasonable performance. If the ovs-vswitchd log file indicates that
227 a feature is not supported, consider upgrading to a newer upstream
228 Linux release or using the kernel module paired with the userspace
231 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
233 A: Support for tunnels was added to the upstream Linux kernel module
234 after the rest of Open vSwitch. As a result, some kernels may contain
235 support for Open vSwitch but not tunnels. The minimum kernel version
236 that supports each tunnel protocol is:
238 | Protocol | Linux Kernel
239 |:--------:|:-------------:
243 | LISP | <not upstream>
244 | STT | <not upstream>
246 If you are using a version of the kernel that is older than the one
247 listed above, it is still possible to use that tunnel protocol. However,
248 you must compile and install the kernel module included with the Open
249 vSwitch distribution rather than the one on your machine. If problems
250 persist after doing this, check to make sure that the module that is
251 loaded is the one you expect.
253 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
255 A: Generating outer UDP checksums requires kernel support that was not
256 part of the initial implementation of these protocols. If using the
257 upstream Linux Open vSwitch module, you must use kernel 4.0 or
258 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
259 support UDP checksums.
261 ### Q: What features are not available when using the userspace datapath?
263 A: Tunnel virtual ports are not supported, as described in the
264 previous answer. It is also not possible to use queue-related
265 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
266 may not be transmitted.
268 ### Q: What Linux kernel versions does IPFIX flow monitoring work with?
270 A: IPFIX flow monitoring requires the Linux kernel module from Linux
271 3.10 or later, or the out-of-tree module from Open vSwitch version
274 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
276 In general, the Open vSwitch userspace should be used with the
277 kernel version included in the same release or with the version
278 from upstream Linux. However, when upgrading between two releases
279 of Open vSwitch it is best to migrate userspace first to reduce
280 the possibility of incompatibilities.
282 ### Q: What happened to the bridge compatibility feature?
284 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
285 When it was enabled, Open vSwitch imitated the interface of the
286 Linux kernel "bridge" module. This allowed users to drop Open
287 vSwitch into environments designed to use the Linux kernel bridge
288 module without adapting the environment to use Open vSwitch.
290 Open vSwitch 1.10 and later do not support bridge compatibility.
291 The feature was dropped because version 1.10 adopted a new internal
292 architecture that made bridge compatibility difficult to maintain.
293 Now that many environments use OVS directly, it would be rarely
296 To use bridge compatibility, install OVS 1.9 or earlier, including
297 the accompanying kernel modules (both the main and bridge
298 compatibility modules), following the instructions that come with
299 the release. Be sure to start the ovs-brcompatd daemon.
305 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
307 A: In networking, the terms "bridge" and "switch" are synonyms. Open
308 vSwitch implements an Ethernet switch, which means that it is also
311 ### Q: What's a VLAN?
313 A: See the "VLAN" section below.
319 ### Q: How do I configure a port as an access port?
321 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
322 the following commands configure br0 with eth0 as a trunk port (the
323 default) and tap0 as an access port for VLAN 9:
326 ovs-vsctl add-port br0 eth0
327 ovs-vsctl add-port br0 tap0 tag=9
329 If you want to configure an already added port as an access port,
330 use "ovs-vsctl set", e.g.:
332 ovs-vsctl set port tap0 tag=9
334 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
336 A: The following commands configure br0 with eth0 and tap0 as trunk
337 ports. All traffic coming in or going out on eth0 or tap0 is also
338 mirrored to tap1; any traffic arriving on tap1 is dropped:
341 ovs-vsctl add-port br0 eth0
342 ovs-vsctl add-port br0 tap0
343 ovs-vsctl add-port br0 tap1 \
344 -- --id=@p get port tap1 \
345 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
346 -- set bridge br0 mirrors=@m
348 To later disable mirroring, run:
350 ovs-vsctl clear bridge br0 mirrors
352 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
354 A: Yes. How you configure it depends on what you mean by "promiscuous
357 - Conventionally, "promiscuous mode" is a feature of a network
358 interface card. Ordinarily, a NIC passes to the CPU only the
359 packets actually destined to its host machine. It discards
360 the rest to avoid wasting memory and CPU cycles. When
361 promiscuous mode is enabled, however, it passes every packet
362 to the CPU. On an old-style shared-media or hub-based
363 network, this allows the host to spy on all packets on the
364 network. But in the switched networks that are almost
365 everywhere these days, promiscuous mode doesn't have much
366 effect, because few packets not destined to a host are
367 delivered to the host's NIC.
369 This form of promiscuous mode is configured in the guest OS of
370 the VMs on your bridge, e.g. with "ifconfig".
372 - The VMware vSwitch uses a different definition of "promiscuous
373 mode". When you configure promiscuous mode on a VMware vNIC,
374 the vSwitch sends a copy of every packet received by the
375 vSwitch to that vNIC. That has a much bigger effect than just
376 enabling promiscuous mode in a guest OS. Rather than getting
377 a few stray packets for which the switch does not yet know the
378 correct destination, the vNIC gets every packet. The effect
379 is similar to replacing the vSwitch by a virtual hub.
381 This "promiscuous mode" is what switches normally call "port
382 mirroring" or "SPAN". For information on how to configure
383 SPAN, see "How do I configure a port as a SPAN port, that is,
384 enable mirroring of all traffic to that port?"
386 ### Q: How do I configure a DPDK port as an access port?
388 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
390 If your version is DPDK-enabled it will support the --dpdk
391 argument on the command line and will display lines with
392 "EAL:..." during startup when --dpdk is supplied.
394 Secondly, when adding a DPDK port, unlike a system port, the
395 type for the interface must be specified. For example;
398 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
400 Finally, it is required that DPDK port names begin with 'dpdk'.
402 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
405 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
407 A: The following commands configure br0 with eth0 as a trunk port and
408 tap0 as an access port for VLAN 10. All traffic coming in or going
409 out on tap0, as well as traffic coming in or going out on eth0 in
410 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
411 VLAN 10, in cases where one is present, is dropped as part of
415 ovs-vsctl add-port br0 eth0
416 ovs-vsctl add-port br0 tap0 tag=10
418 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
420 -- set bridge br0 mirrors=@m
422 To later disable mirroring, run:
424 ovs-vsctl clear bridge br0 mirrors
426 Mirroring to a VLAN can disrupt a network that contains unmanaged
427 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
428 GRE tunnel has fewer caveats than mirroring to a VLAN and should
429 generally be preferred.
431 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
433 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
434 of the specified output-vlan. This loss of information may make
435 the mirrored traffic too hard to interpret.
437 To mirror multiple VLANs, use the commands above, but specify a
438 comma-separated list of VLANs as the value for select-vlan. To
439 mirror every VLAN, use the commands above, but omit select-vlan and
442 When a packet arrives on a VLAN that is used as a mirror output
443 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
444 floods the packet across all the ports for which the mirror output
445 VLAN is configured. (If an OpenFlow controller is in use, then it
446 can override this behavior through the flow table.) If OVS is used
447 as an intermediate switch, rather than an edge switch, this ensures
448 that the RSPAN traffic is distributed through the network.
450 Mirroring to a VLAN can disrupt a network that contains unmanaged
451 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
452 GRE tunnel has fewer caveats than mirroring to a VLAN and should
453 generally be preferred.
455 ### Q: How do I configure mirroring of all traffic to a GRE tunnel?
457 A: The following commands configure br0 with eth0 and tap0 as trunk
458 ports. All traffic coming in or going out on eth0 or tap0 is also
459 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
460 traffic arriving on gre0 is dropped:
463 ovs-vsctl add-port br0 eth0
464 ovs-vsctl add-port br0 tap0
465 ovs-vsctl add-port br0 gre0 \
466 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
467 -- --id=@p get port gre0 \
468 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
469 -- set bridge br0 mirrors=@m
471 To later disable mirroring and destroy the GRE tunnel:
473 ovs-vsctl clear bridge br0 mirrors
474 ovs-vsctl del-port br0 gre0
476 ### Q: Does Open vSwitch support ERSPAN?
478 A: No. ERSPAN is an undocumented proprietary protocol. As an
479 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
482 ### Q: How do I connect two bridges?
484 A: First, why do you want to do this? Two connected bridges are not
485 much different from a single bridge, so you might as well just have
486 a single bridge with all your ports on it.
488 If you still want to connect two bridges, you can use a pair of
489 patch ports. The following example creates bridges br0 and br1,
490 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
491 and br1 with a pair of patch ports.
494 ovs-vsctl add-port br0 eth0
495 ovs-vsctl add-port br0 tap0
497 ovs-vsctl add-port br1 tap1
499 -- add-port br0 patch0 \
500 -- set interface patch0 type=patch options:peer=patch1 \
501 -- add-port br1 patch1 \
502 -- set interface patch1 type=patch options:peer=patch0
504 Bridges connected with patch ports are much like a single bridge.
505 For instance, if the example above also added eth1 to br1, and both
506 eth0 and eth1 happened to be connected to the same next-hop switch,
507 then you could loop your network just as you would if you added
508 eth0 and eth1 to the same bridge (see the "Configuration Problems"
509 section below for more information).
511 If you are using Open vSwitch 1.9 or an earlier version, then you
512 need to be using the kernel module bundled with Open vSwitch rather
513 than the one that is integrated into Linux 3.3 and later, because
514 Open vSwitch 1.9 and earlier versions need kernel support for patch
515 ports. This also means that in Open vSwitch 1.9 and earlier, patch
516 ports will not work with the userspace datapath, only with the
519 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
521 A: Open vSwitch does not support such a configuration.
522 Bridges always have their local ports.
525 Implementation Details
526 ----------------------
528 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
530 A: Open vSwitch uses different kinds of flows for different purposes:
532 - OpenFlow flows are the most important kind of flow. OpenFlow
533 controllers use these flows to define a switch's policy.
534 OpenFlow flows support wildcards, priorities, and multiple
537 When in-band control is in use, Open vSwitch sets up a few
538 "hidden" flows, with priority higher than a controller or the
539 user can configure, that are not visible via OpenFlow. (See
540 the "Controller" section of the FAQ for more information
543 - The Open vSwitch software switch implementation uses a second
544 kind of flow internally. These flows, called "datapath" or
545 "kernel" flows, do not support priorities and comprise only a
546 single table, which makes them suitable for caching. (Like
547 OpenFlow flows, datapath flows do support wildcarding, in Open
548 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
549 also support different actions and number ports differently.
551 Datapath flows are an implementation detail that is subject to
552 change in future versions of Open vSwitch. Even with the
553 current version of Open vSwitch, hardware switch
554 implementations do not necessarily use this architecture.
556 Users and controllers directly control only the OpenFlow flow
557 table. Open vSwitch manages the datapath flow table itself, so
558 users should not normally be concerned with it.
560 ### Q: Why are there so many different ways to dump flows?
562 A: Open vSwitch has two kinds of flows (see the previous question), so
563 it has commands with different purposes for dumping each kind of
566 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
567 hidden flows. This is the most commonly useful form of flow
568 dump. (Unlike the other commands, this should work with any
569 OpenFlow switch, not just Open vSwitch.)
571 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
572 including hidden flows. This is occasionally useful for
573 troubleshooting suspected issues with in-band control.
575 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
576 entries for a Linux kernel-based datapath. In Open vSwitch
577 1.10 and later, ovs-vswitchd merges multiple switches into a
578 single datapath, so it will show all the flows on all your
579 kernel-based switches. This command can occasionally be
580 useful for debugging.
582 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
583 dumps datapath flows for only the specified bridge, regardless
586 ### Q: How does multicast snooping works with VLANs?
588 A: Open vSwitch maintains snooping tables for each VLAN.
590 ### Q: Can OVS populate the kernel flow table in advance instead of in reaction to packets?
592 A: No. There are several reasons:
594 - Kernel flows are not as sophisticated as OpenFlow flows, which
595 means that some OpenFlow policies could require a large number of
596 kernel flows. The "conjunctive match" feature is an extreme
597 example: the number of kernel flows it requires is the product of
598 the number of flows in each dimension.
600 - With multiple OpenFlow flow tables and simple sets of actions, the
601 number of kernel flows required can be as large as the product of
602 the number of flows in each dimension. With more sophisticated
603 actions, the number of kernel flows could be even larger.
605 - Open vSwitch is designed so that any version of OVS userspace
606 interoperates with any version of the OVS kernel module. This
607 forward and backward compatibility requires that userspace observe
608 how the kernel module parses received packets. This is only
609 possible in a straightforward way when userspace adds kernel flows
610 in reaction to received packets.
612 For more relevant information on the architecture of Open vSwitch,
613 please read "The Design and Implementation of Open vSwitch",
614 published in USENIX NSDI 2015.
620 ### Q: I just upgraded and I see a performance drop. Why?
622 A: The OVS kernel datapath may have been updated to a newer version than
623 the OVS userspace components. Sometimes new versions of OVS kernel
624 module add functionality that is backwards compatible with older
625 userspace components but may cause a drop in performance with them.
626 Especially, if a kernel module from OVS 2.1 or newer is paired with
627 OVS userspace 1.10 or older, there will be a performance drop for
630 Updating the OVS userspace components to the latest released
631 version should fix the performance degradation.
633 To get the best possible performance and functionality, it is
634 recommended to pair the same versions of the kernel module and OVS
638 Configuration Problems
639 ----------------------
641 ### Q: I created a bridge and added my Ethernet port to it, using commands
645 ovs-vsctl add-port br0 eth0
647 and as soon as I ran the "add-port" command I lost all connectivity
650 A: A physical Ethernet device that is part of an Open vSwitch bridge
651 should not have an IP address. If one does, then that IP address
652 will not be fully functional.
654 You can restore functionality by moving the IP address to an Open
655 vSwitch "internal" device, such as the network device named after
656 the bridge itself. For example, assuming that eth0's IP address is
657 192.168.128.5, you could run the commands below to fix up the
660 ifconfig eth0 0.0.0.0
661 ifconfig br0 192.168.128.5
663 (If your only connection to the machine running OVS is through the
664 IP address in question, then you would want to run all of these
665 commands on a single command line, or put them into a script.) If
666 there were any additional routes assigned to eth0, then you would
667 also want to use commands to adjust these routes to go through br0.
669 If you use DHCP to obtain an IP address, then you should kill the
670 DHCP client that was listening on the physical Ethernet interface
671 (e.g. eth0) and start one listening on the internal interface
672 (e.g. br0). You might still need to manually clear the IP address
673 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
675 There is no compelling reason why Open vSwitch must work this way.
676 However, this is the way that the Linux kernel bridge module has
677 always worked, so it's a model that those accustomed to Linux
678 bridging are already used to. Also, the model that most people
679 expect is not implementable without kernel changes on all the
680 versions of Linux that Open vSwitch supports.
682 By the way, this issue is not specific to physical Ethernet
683 devices. It applies to all network devices except Open vSwitch
686 ### Q: I created a bridge and added a couple of Ethernet ports to it,
687 ### using commands like these:
690 ovs-vsctl add-port br0 eth0
691 ovs-vsctl add-port br0 eth1
693 and now my network seems to have melted: connectivity is unreliable
694 (even connectivity that doesn't go through Open vSwitch), all the
695 LEDs on my physical switches are blinking, wireshark shows
696 duplicated packets, and CPU usage is very high.
698 A: More than likely, you've looped your network. Probably, eth0 and
699 eth1 are connected to the same physical Ethernet switch. This
700 yields a scenario where OVS receives a broadcast packet on eth0 and
701 sends it out on eth1, then the physical switch connected to eth1
702 sends the packet back on eth0, and so on forever. More complicated
703 scenarios, involving a loop through multiple switches, are possible
706 The solution depends on what you are trying to do:
708 - If you added eth0 and eth1 to get higher bandwidth or higher
709 reliability between OVS and your physical Ethernet switch,
710 use a bond. The following commands create br0 and then add
711 eth0 and eth1 as a bond:
714 ovs-vsctl add-bond br0 bond0 eth0 eth1
716 Bonds have tons of configuration options. Please read the
717 documentation on the Port table in ovs-vswitchd.conf.db(5)
720 Configuration for DPDK-enabled interfaces is slightly less
721 straightforward: see [INSTALL.DPDK.md].
723 - Perhaps you don't actually need eth0 and eth1 to be on the
724 same bridge. For example, if you simply want to be able to
725 connect each of them to virtual machines, then you can put
726 each of them on a bridge of its own:
729 ovs-vsctl add-port br0 eth0
732 ovs-vsctl add-port br1 eth1
734 and then connect VMs to br0 and br1. (A potential
735 disadvantage is that traffic cannot directly pass between br0
736 and br1. Instead, it will go out eth0 and come back in eth1,
739 - If you have a redundant or complex network topology and you
740 want to prevent loops, turn on spanning tree protocol (STP).
741 The following commands create br0, enable STP, and add eth0
742 and eth1 to the bridge. The order is important because you
743 don't want have to have a loop in your network even
747 ovs-vsctl set bridge br0 stp_enable=true
748 ovs-vsctl add-port br0 eth0
749 ovs-vsctl add-port br0 eth1
751 The Open vSwitch implementation of STP is not well tested.
752 Please report any bugs you observe, but if you'd rather avoid
753 acting as a beta tester then another option might be your
756 ### Q: I can't seem to use Open vSwitch in a wireless network.
758 A: Wireless base stations generally only allow packets with the source
759 MAC address of NIC that completed the initial handshake.
760 Therefore, without MAC rewriting, only a single device can
761 communicate over a single wireless link.
763 This isn't specific to Open vSwitch, it's enforced by the access
764 point, so the same problems will show up with the Linux bridge or
765 any other way to do bridging.
767 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
769 A: PPP most commonly carries IP packets, but Open vSwitch works only
770 with Ethernet frames. The correct way to interface PPP to an
771 Ethernet network is usually to use routing instead of switching.
773 ### Q: Is there any documentation on the database tables and fields?
775 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
777 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
778 I only see a datapath called "ovs-system". How can I see datapath
779 information about a particular bridge?
781 A: In version 1.9.0, OVS switched to using a single datapath that is
782 shared by all bridges of that type. The "ovs-appctl dpif/*"
783 commands provide similar functionality that is scoped by the bridge.
785 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
786 see the port in the datapath?
788 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
789 cannot be loaded at the same time. It is likely that on your system the
790 Linux GRE module is already loaded and blocking OVS (to confirm, check
791 dmesg for errors regarding GRE registration). To fix this, unload all
792 GRE modules that appear in lsmod as well as the OVS kernel module. You
793 can then reload the OVS module following the directions in
794 [INSTALL.md], which will ensure that dependencies are satisfied.
796 ### Q: Open vSwitch does not seem to obey my packet filter rules.
798 A: It depends on mechanisms and configurations you want to use.
800 You cannot usefully use typical packet filters, like iptables, on
801 physical Ethernet ports that you add to an Open vSwitch bridge.
802 This is because Open vSwitch captures packets from the interface at
803 a layer lower below where typical packet-filter implementations
804 install their hooks. (This actually applies to any interface of
805 type "system" that you might add to an Open vSwitch bridge.)
807 You can usefully use typical packet filters on Open vSwitch
808 internal ports as they are mostly ordinary interfaces from the point
809 of view of packet filters.
811 For example, suppose you create a bridge br0 and add Ethernet port
812 eth0 to it. Then you can usefully add iptables rules to affect the
813 internal interface br0, but not the physical interface eth0. (br0
814 is also where you would add an IP address, as discussed elsewhere
817 For simple filtering rules, it might be possible to achieve similar
818 results by installing appropriate OpenFlow flows instead.
820 If the use of a particular packet filter setup is essential, Open
821 vSwitch might not be the best choice for you. On Linux, you might
822 want to consider using the Linux Bridge. (This is the only choice if
823 you want to use ebtables rules.) On NetBSD, you might want to
824 consider using the bridge(4) with BRIDGE_IPF option.
826 ### Q: It seems that Open vSwitch does nothing when I removed a port and
827 then immediately put it back. For example, consider that p1 is
828 a port of type=internal:
830 ovs-vsctl del-port br0 p1 -- \
832 set interface p1 type=internal
834 A: It's an expected behaviour.
836 If del-port and add-port happen in a single OVSDB transaction as
837 your example, Open vSwitch always "skips" the intermediate steps.
838 Even if they are done in multiple transactions, it's still allowed
839 for Open vSwitch to skip the intermediate steps and just implement
840 the overall effect. In both cases, your example would be turned
843 If you want to make Open vSwitch actually destroy and then re-create
844 the port for some side effects like resetting kernel setting for the
845 corresponding interface, you need to separate operations into multiple
846 OVSDB transactions and ensure that at least the first one does not have
847 --no-wait. In the following example, the first ovs-vsctl will block
848 until Open vSwitch reloads the new configuration and removes the port:
850 ovs-vsctl del-port br0 p1
851 ovs-vsctl add-port br0 p1 -- \
852 set interface p1 type=internal
854 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
855 it takes too long (minutes or hours) to do it with ovs-vsctl. How
858 A: If you add them one at a time with ovs-vsctl, it can take a long
859 time to add thousands of ports to an Open vSwitch bridge. This is
860 because every invocation of ovs-vsctl first reads the current
861 configuration from OVSDB. As the number of ports grows, this
862 starts to take an appreciable amount of time, and when it is
863 repeated thousands of times the total time becomes significant.
865 The solution is to add the ports in one invocation of ovs-vsctl (or
866 a small number of them). For example, using bash:
869 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
872 takes seconds, not minutes or hours, in the OVS sandbox environment.
874 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
875 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
878 A: Open vSwitch wasn't able to create the bridge. Check the
879 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
880 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
882 In general, the Open vSwitch database reflects the desired
883 configuration state. ovs-vswitchd monitors the database and, when
884 it changes, reconfigures the system to reflect the new desired
885 state. This normally happens very quickly. Thus, a discrepancy
886 between the database and the actual state indicates that
887 ovs-vswitchd could not implement the configuration, and so one
888 should check the log to find out why. (Another possible cause is
889 that ovs-vswitchd is not running. This will make "ovs-vsctl"
890 commands hang, if they change the configuration, unless one
891 specifies "--no-wait".)
893 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
894 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
895 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
898 A: Open vSwitch wasn't able to create the port. Check the
899 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
900 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
901 see the previous question for more information.
903 You may want to upgrade to Open vSwitch 2.3 (or later), in which
904 ovs-vsctl will immediately report when there is an issue creating a
907 ### Q: I created a tap device tap0, configured an IP address on it, and
908 added it to a bridge, like this:
911 ifconfig tap0 192.168.0.123
913 ovs-vsctl add-port br0 tap0
915 I expected that I could then use this IP address to contact other
916 hosts on the network, but it doesn't work. Why not?
918 A: The short answer is that this is a misuse of a "tap" device. Use
919 an "internal" device implemented by Open vSwitch, which works
920 differently and is designed for this use. To solve this problem
921 with an internal device, instead run:
924 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
925 ifconfig int0 192.168.0.123
927 Even more simply, you can take advantage of the internal port that
928 every bridge has under the name of the bridge:
931 ifconfig br0 192.168.0.123
933 In more detail, a "tap" device is an interface between the Linux
934 (or *BSD) network stack and a user program that opens it as a
935 socket. When the "tap" device transmits a packet, it appears in
936 the socket opened by the userspace program. Conversely, when the
937 userspace program writes to the "tap" socket, the kernel TCP/IP
938 stack processes the packet as if it had been received by the "tap"
941 Consider the configuration above. Given this configuration, if you
942 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
943 routing stack will transmit an ARP on the tap0 device. Open
944 vSwitch userspace treats "tap" devices just like any other network
945 device; that is, it doesn't open them as "tap" sockets. That means
946 that the ARP packet will simply get dropped.
948 You might wonder why the Open vSwitch kernel module doesn't
949 intercept the ARP packet and bridge it. After all, Open vSwitch
950 intercepts packets on other devices. The answer is that Open
951 vSwitch only intercepts *received* packets, but this is a packet
952 being transmitted. The same thing happens for all other types of
953 network devices, except for Open vSwitch "internal" ports. If you,
954 for example, add a physical Ethernet port to an OVS bridge,
955 configure an IP address on a physical Ethernet port, and then issue
956 a "ping" to an address in that subnet, the same thing happens: an
957 ARP gets transmitted on the physical Ethernet port and Open vSwitch
958 never sees it. (You should not do that, as documented at the
959 beginning of this section.)
961 It can make sense to add a "tap" device to an Open vSwitch bridge,
962 if some userspace program (other than Open vSwitch) has opened the
963 tap socket. This is the case, for example, if the "tap" device was
964 created by KVM (or QEMU) to simulate a virtual NIC. In such a
965 case, when OVS bridges a packet to the "tap" device, the kernel
966 forwards that packet to KVM in userspace, which passes it along to
967 the VM, and in the other direction, when the VM sends a packet, KVM
968 writes it to the "tap" socket, which causes OVS to receive it and
969 bridge it to the other OVS ports. Please note that in such a case
970 no IP address is configured on the "tap" device (there is normally
971 an IP address configured in the virtual NIC inside the VM, but this
972 is not visible to the host Linux kernel or to Open vSwitch).
974 There is one special case in which Open vSwitch does directly read
975 and write "tap" sockets. This is an implementation detail of the
976 Open vSwitch userspace switch, which implements its "internal"
977 ports as Linux (or *BSD) "tap" sockets. In such a userspace
978 switch, OVS receives packets sent on the "tap" device used to
979 implement an "internal" port by reading the associated "tap"
980 socket, and bridges them to the rest of the switch. In the other
981 direction, OVS transmits packets bridged to the "internal" port by
982 writing them to the "tap" socket, causing them to be processed by
983 the kernel TCP/IP stack as if they had been received on the "tap"
984 device. Users should not need to be concerned with this
985 implementation detail.
987 Open vSwitch has a network device type called "tap". This is
988 intended only for implementing "internal" ports in the OVS
989 userspace switch and should not be used otherwise. In particular,
990 users should not configure KVM "tap" devices as type "tap" (use
991 type "system", the default, instead).
994 Quality of Service (QoS)
995 ------------------------
997 ### Q: Does OVS support Quality of Service (QoS)?
999 A: Yes. For traffic that egresses from a switch, OVS supports traffic
1000 shaping; for traffic that ingresses into a switch, OVS support
1001 policing. Policing is a simple form of quality-of-service that
1002 simply drops packets received in excess of the configured rate. Due
1003 to its simplicity, policing is usually less accurate and less
1004 effective than egress traffic shaping, which queues packets.
1006 Keep in mind that ingress and egress are from the perspective of the
1007 switch. That means that egress shaping limits the rate at which
1008 traffic is allowed to transmit from a physical interface, but the
1009 rate at which traffic will be received on a virtual machine's VIF.
1010 For ingress policing, the behavior is the opposite.
1012 ### Q: How do I configure egress traffic shaping?
1014 A: Suppose that you want to set up bridge br0 connected to physical
1015 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
1016 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
1017 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
1018 could configure the bridge this way:
1022 add-port br0 eth0 -- \
1023 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
1024 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
1025 set port eth0 qos=@newqos -- \
1026 --id=@newqos create qos type=linux-htb \
1027 other-config:max-rate=1000000000 \
1028 queues:123=@vif10queue \
1029 queues:234=@vif20queue -- \
1030 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
1031 --id=@vif20queue create queue other-config:max-rate=20000000
1033 At this point, bridge br0 is configured with the ports and eth0 is
1034 configured with the queues that you need for QoS, but nothing is
1035 actually directing packets from vif1.0 or vif2.0 to the queues that
1036 we have set up for them. That means that all of the packets to
1037 eth0 are going to the "default queue", which is not what we want.
1039 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
1040 queues reserved for them:
1042 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
1043 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
1045 Each of the above flows matches on the input port, sets up the
1046 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
1047 executes the "normal" action, which performs the same switching
1048 that Open vSwitch would have done without any OpenFlow flows being
1049 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1050 numbers 5 and 6, respectively, because we set their ofport_request
1051 columns above. If we had not done that, then we would have needed
1052 to find out their port numbers before setting up these flows.)
1054 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1057 By the way, if you delete the bridge created by the above commands,
1060 ovs-vsctl del-br br0
1062 then that will leave one unreferenced QoS record and two
1063 unreferenced Queue records in the Open vSwich database. One way to
1064 clear them out, assuming you don't have other QoS or Queue records
1065 that you want to keep, is:
1067 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1069 If you do want to keep some QoS or Queue records, or the Open
1070 vSwitch you are using is older than version 1.8 (which added the
1071 --all option), then you will have to destroy QoS and Queue records
1074 ### Q: How do I configure ingress policing?
1076 A: A policing policy can be configured on an interface to drop packets
1077 that arrive at a higher rate than the configured value. For example,
1078 the following commands will rate-limit traffic that vif1.0 may
1081 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1082 ovs-vsctl set interface vif1.0 ingress_policing_burst=1000
1084 Traffic policing can interact poorly with some network protocols and
1085 can have surprising results. The "Ingress Policing" section of
1086 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1088 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1089 adding records to the QoS and Queue table, but the results aren't
1092 A: Did you install OpenFlow flows that use your queues? This is the
1093 primary way to tell Open vSwitch which queues you want to use. If
1094 you don't do this, then the default queue will be used, which will
1095 probably not have the effect you want.
1097 Refer to the previous question for an example.
1099 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1100 doesn't yet support. How do I do that?
1102 A: Open vSwitch does not implement QoS itself. Instead, it can
1103 configure some, but not all, of the QoS features built into the
1104 Linux kernel. If you need some QoS feature that OVS cannot
1105 configure itself, then the first step is to figure out whether
1106 Linux QoS supports that feature. If it does, then you can submit a
1107 patch to support Open vSwitch configuration for that feature, or
1108 you can use "tc" directly to configure the feature in Linux. (If
1109 Linux QoS doesn't support the feature you want, then first you have
1110 to add that support to Linux.)
1112 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1113 working as well as I expect.
1115 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1118 - Open vSwitch configures a subset of Linux kernel QoS
1119 features, according to what is in OVSDB. It is possible that
1120 this code has bugs. If you believe that this is so, then you
1121 can configure the Linux traffic control (QoS) stack directly
1122 with the "tc" program. If you get better results that way,
1123 you can send a detailed bug report to bugs@openvswitch.org.
1125 It is certain that Open vSwitch cannot configure every Linux
1126 kernel QoS feature. If you need some feature that OVS cannot
1127 configure, then you can also use "tc" directly (or add that
1130 - The Open vSwitch implementation of OpenFlow allows flows to
1131 be directed to particular queues. This is pretty simple and
1132 unlikely to have serious bugs at this point.
1134 However, most problems with QoS on Linux are not bugs in Open
1135 vSwitch at all. They tend to be either configuration errors
1136 (please see the earlier questions in this section) or issues with
1137 the traffic control (QoS) stack in Linux. The Open vSwitch
1138 developers are not experts on Linux traffic control. We suggest
1139 that, if you believe you are encountering a problem with Linux
1140 traffic control, that you consult the tc manpages (e.g. tc(8),
1141 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1142 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1144 ### Q: Does Open vSwitch support OpenFlow meters?
1146 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1147 OpenFlow meters. There is no implementation of meters in the Open
1148 vSwitch software switch (neither the kernel-based nor userspace
1155 ### Q: What's a VLAN?
1157 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1158 partition a single switch into multiple switches. Suppose, for
1159 example, that you have two groups of machines, group A and group B.
1160 You want the machines in group A to be able to talk to each other,
1161 and you want the machine in group B to be able to talk to each
1162 other, but you don't want the machines in group A to be able to
1163 talk to the machines in group B. You can do this with two
1164 switches, by plugging the machines in group A into one switch and
1165 the machines in group B into the other switch.
1167 If you only have one switch, then you can use VLANs to do the same
1168 thing, by configuring the ports for machines in group A as VLAN
1169 "access ports" for one VLAN and the ports for group B as "access
1170 ports" for a different VLAN. The switch will only forward packets
1171 between ports that are assigned to the same VLAN, so this
1172 effectively subdivides your single switch into two independent
1173 switches, one for each group of machines.
1175 So far we haven't said anything about VLAN headers. With access
1176 ports, like we've described so far, no VLAN header is present in
1177 the Ethernet frame. This means that the machines (or switches)
1178 connected to access ports need not be aware that VLANs are
1179 involved, just like in the case where we use two different physical
1182 Now suppose that you have a whole bunch of switches in your
1183 network, instead of just one, and that some machines in group A are
1184 connected directly to both switches 1 and 2. To allow these
1185 machines to talk to each other, you could add an access port for
1186 group A's VLAN to switch 1 and another to switch 2, and then
1187 connect an Ethernet cable between those ports. That works fine,
1188 but it doesn't scale well as the number of switches and the number
1189 of VLANs increases, because you use up a lot of valuable switch
1190 ports just connecting together your VLANs.
1192 This is where VLAN headers come in. Instead of using one cable and
1193 two ports per VLAN to connect a pair of switches, we configure a
1194 port on each switch as a VLAN "trunk port". Packets sent and
1195 received on a trunk port carry a VLAN header that says what VLAN
1196 the packet belongs to, so that only two ports total are required to
1197 connect the switches, regardless of the number of VLANs in use.
1198 Normally, only switches (either physical or virtual) are connected
1199 to a trunk port, not individual hosts, because individual hosts
1200 don't expect to see a VLAN header in the traffic that they receive.
1202 None of the above discussion says anything about particular VLAN
1203 numbers. This is because VLAN numbers are completely arbitrary.
1204 One must only ensure that a given VLAN is numbered consistently
1205 throughout a network and that different VLANs are given different
1206 numbers. (That said, VLAN 0 is usually synonymous with a packet
1207 that has no VLAN header, and VLAN 4095 is reserved.)
1209 ### Q: VLANs don't work.
1211 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1212 bugs. If you are having problems with VLANs that you suspect to be
1213 driver related, then you have several options:
1215 - Upgrade to Linux 3.3 or later.
1217 - Build and install a fixed version of the particular driver
1218 that is causing trouble, if one is available.
1220 - Use a NIC whose driver does not have VLAN problems.
1222 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
1223 that works around bugs in kernel drivers. To enable VLAN
1224 splinters on interface eth0, use the command:
1226 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1228 For VLAN splinters to be effective, Open vSwitch must know
1229 which VLANs are in use. See the "VLAN splinters" section in
1230 the Interface table in ovs-vswitchd.conf.db(5) for details on
1231 how Open vSwitch infers in-use VLANs.
1233 VLAN splinters increase memory use and reduce performance, so
1234 use them only if needed.
1236 - Apply the "vlan workaround" patch from the XenServer kernel
1237 patch queue, build Open vSwitch against this patched kernel,
1238 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1239 workaround for each interface whose driver is buggy.
1241 (This is a nontrivial exercise, so this option is included
1242 only for completeness.)
1244 It is not always easy to tell whether a Linux kernel driver has
1245 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1246 can help you test. See their manpages for details. Of the two
1247 utilities, ovs-test(8) is newer and more thorough, but
1248 ovs-vlan-test(8) may be easier to use.
1250 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1252 A: Do you have VLANs enabled on the physical switch that OVS is
1253 attached to? Make sure that the port is configured to trunk the
1254 VLAN or VLANs that you are using with OVS.
1256 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1257 and to its destination host, but OVS seems to drop incoming return
1260 A: It's possible that you have the VLAN configured on your physical
1261 switch as the "native" VLAN. In this mode, the switch treats
1262 incoming packets either tagged with the native VLAN or untagged as
1263 part of the native VLAN. It may also send outgoing packets in the
1264 native VLAN without a VLAN tag.
1266 If this is the case, you have two choices:
1268 - Change the physical switch port configuration to tag packets
1269 it forwards to OVS with the native VLAN instead of forwarding
1272 - Change the OVS configuration for the physical port to a
1273 native VLAN mode. For example, the following sets up a
1274 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1276 ovs-vsctl add-br br0
1277 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1279 In this situation, "native-untagged" mode will probably work
1280 equally well. Refer to the documentation for the Port table
1281 in ovs-vswitchd.conf.db(5) for more information.
1283 ### Q: I added a pair of VMs on different VLANs, like this:
1285 ovs-vsctl add-br br0
1286 ovs-vsctl add-port br0 eth0
1287 ovs-vsctl add-port br0 tap0 tag=9
1288 ovs-vsctl add-port br0 tap1 tag=10
1290 but the VMs can't access each other, the external network, or the
1293 A: It is to be expected that the VMs can't access each other. VLANs
1294 are a means to partition a network. When you configured tap0 and
1295 tap1 as access ports for different VLANs, you indicated that they
1296 should be isolated from each other.
1298 As for the external network and the Internet, it seems likely that
1299 the machines you are trying to access are not on VLAN 9 (or 10) and
1300 that the Internet is not available on VLAN 9 (or 10).
1302 ### Q: I added a pair of VMs on the same VLAN, like this:
1304 ovs-vsctl add-br br0
1305 ovs-vsctl add-port br0 eth0
1306 ovs-vsctl add-port br0 tap0 tag=9
1307 ovs-vsctl add-port br0 tap1 tag=9
1309 The VMs can access each other, but not the external network or the
1312 A: It seems likely that the machines you are trying to access in the
1313 external network are not on VLAN 9 and that the Internet is not
1314 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1315 trunk VLAN on the upstream switch port to which eth0 is connected.
1317 ### Q: Can I configure an IP address on a VLAN?
1319 A: Yes. Use an "internal port" configured as an access port. For
1320 example, the following configures IP address 192.168.0.7 on VLAN 9.
1321 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1322 they have an 802.1Q header with VLAN 9. Conversely, traffic
1323 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1326 ovs-vsctl add-br br0
1327 ovs-vsctl add-port br0 eth0
1328 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1329 ifconfig vlan9 192.168.0.7
1331 See also the following question.
1333 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1336 ovs-vsctl add-br br0
1337 ovs-vsctl add-port br0 eth0
1338 ifconfig br0 192.168.0.5
1339 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1340 ifconfig vlan9 192.168.0.9
1342 but other hosts that are only on VLAN 0 can reach the IP address
1343 configured on VLAN 9. What's going on?
1345 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1346 approaches to IP address handling in Internet hosts:
1348 - In the "Strong ES Model", where an ES is a host ("End
1349 System"), an IP address is primarily associated with a
1350 particular interface. The host discards packets that arrive
1351 on interface A if they are destined for an IP address that is
1352 configured on interface B. The host never sends packets from
1353 interface A using a source address configured on interface B.
1355 - In the "Weak ES Model", an IP address is primarily associated
1356 with a host. The host accepts packets that arrive on any
1357 interface if they are destined for any of the host's IP
1358 addresses, even if the address is configured on some
1359 interface other than the one on which it arrived. The host
1360 does not restrict itself to sending packets from an IP
1361 address associated with the originating interface.
1363 Linux uses the weak ES model. That means that when packets
1364 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1365 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1366 address, even though they were not received on vlan9, the network
1369 To simulate the strong ES model on Linux, one may add iptables rule
1370 to filter packets based on source and destination address and
1371 adjust ARP configuration with sysctls.
1373 BSD uses the strong ES model.
1375 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1377 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1378 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1379 implementation of the OpenFlow "normal switching" action. By
1380 default, when Open vSwitch isn't connected to a controller and
1381 nothing has been manually configured in the flow table, all traffic
1382 goes through the "normal switching" action. But, if you set up
1383 OpenFlow flows on your own, through a controller or using ovs-ofctl
1384 or through other means, then you have to implement VLAN handling
1387 You can use "normal switching" as a component of your OpenFlow
1388 actions, e.g. by putting "normal" into the lists of actions on
1389 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1390 controller. In situations where this is not suitable, you can
1391 implement VLAN handling yourself, e.g.:
1393 - If a packet comes in on an access port, and the flow table
1394 needs to send it out on a trunk port, then the flow can add
1395 the appropriate VLAN tag with the "mod_vlan_vid" action.
1397 - If a packet comes in on a trunk port, and the flow table
1398 needs to send it out on an access port, then the flow can
1399 strip the VLAN tag with the "strip_vlan" action.
1401 ### Q: I configured ports on a bridge as access ports with different VLAN
1404 ovs-vsctl add-br br0
1405 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1406 ovs-vsctl add-port br0 eth0
1407 ovs-vsctl add-port br0 tap0 tag=9
1408 ovs-vsctl add-port br0 tap1 tag=10
1410 but the VMs running behind tap0 and tap1 can still communicate,
1411 that is, they are not isolated from each other even though they are
1414 A: Do you have a controller configured on br0 (as the commands above
1415 do)? If so, then this is a variant on the previous question, "My
1416 OpenFlow controller doesn't see the VLANs that I expect," and you
1417 can refer to the answer there for more information.
1419 ### Q: How MAC learning works with VLANs?
1421 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1422 OFPP_NORMAL action. I.e. it logically has separate learning tables
1429 ### Q: What's a VXLAN?
1431 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1432 to solve the scaling challenges of VLAN networks in a multi-tenant
1433 environment. VXLAN is an overlay network which transports an L2 network
1434 over an existing L3 network. For more information on VXLAN, please see
1437 http://tools.ietf.org/html/rfc7348
1439 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1441 A: Open vSwitch currently supports the framing format for packets on the
1442 wire. There is currently no support for the multicast aspects of VXLAN.
1443 To get around the lack of multicast support, it is possible to
1444 pre-provision MAC to IP address mappings either manually or from a
1447 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1450 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1451 is 4789. However, it is possible to configure the destination UDP port
1452 manually on a per-VXLAN tunnel basis. An example of this configuration is
1455 ovs-vsctl add-br br0
1456 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1457 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1458 options:dst_port=8472
1461 Using OpenFlow (Manually or Via Controller)
1462 -------------------------------------------
1464 ### Q: What versions of OpenFlow does Open vSwitch support?
1466 A: The following table lists the versions of OpenFlow supported by
1467 each version of Open vSwitch:
1469 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5
1470 ###============ ===== ===== ===== ===== ===== =====
1471 1.9 and earlier yes --- --- --- --- ---
1472 1.10 yes --- [*] [*] --- ---
1473 1.11 yes --- [*] [*] --- ---
1474 2.0 yes [*] [*] [*] --- ---
1475 2.1 yes [*] [*] [*] --- ---
1476 2.2 yes [*] [*] [*] [%] [*]
1477 2.3 yes yes yes yes [*] [*]
1479 [*] Supported, with one or more missing features.
1480 [%] Experimental, unsafe implementation.
1482 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1483 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1484 1.2, and 1.3 must be enabled manually in ovs-vswitchd. OpenFlow
1485 1.4 and 1.5 are also supported, with missing features, in Open
1486 vSwitch 2.3 and later, but not enabled by default. In any case,
1487 the user may override the default:
1489 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1491 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1493 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1495 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1497 - To enable only OpenFlow 1.0 on bridge br0:
1499 ovs-vsctl set bridge br0 protocols=OpenFlow10
1501 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1502 default. Use the -O option to enable support for later versions of
1503 OpenFlow in ovs-ofctl. For example:
1505 ovs-ofctl -O OpenFlow13 dump-flows br0
1507 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1508 1.4 that could cause crashes. We don't recommend enabling it.)
1510 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1511 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1512 1.5 is solidly implemented, Open vSwitch will enable those version
1515 ### Q: Does Open vSwitch support MPLS?
1517 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1518 these versions can match on MPLS Ethernet types, but they cannot
1519 match, push, or pop MPLS labels, nor can they look past MPLS labels
1520 into the encapsulated packet.
1522 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1523 for MPLS. With the userspace datapath only, these versions can
1524 match, push, or pop a single MPLS label, but they still cannot look
1525 past MPLS labels (even after popping them) into the encapsulated
1526 packet. Kernel datapath support is unchanged from earlier
1529 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1530 label and look past the MPLS label into the encapsulated packet.
1531 Both userspace and kernel datapaths will be supported, but MPLS
1532 processing always happens in userspace either way, so kernel
1533 datapath performance will be disappointing.
1535 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1536 labels and look past the MPLS label into the encapsulated packet.
1537 It will have kernel support for MPLS, yielding improved
1540 ### Q: I'm getting "error type 45250 code 0". What's that?
1542 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1543 vSwitch uses this extension when it must report an error to an
1544 OpenFlow controller but no standard OpenFlow error code is
1547 Open vSwitch logs the errors that it sends to controllers, so the
1548 easiest thing to do is probably to look at the ovs-vswitchd log to
1549 find out what the error was.
1551 If you want to dissect the extended error message yourself, the
1552 format is documented in include/openflow/nicira-ext.h in the Open
1553 vSwitch source distribution. The extended error codes are
1554 documented in lib/ofp-errors.h.
1556 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1557 doesn't actually appear through the OpenFlow connection, even
1558 though I know that it's going through.
1559 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1560 to apply to certain traffic, especially traffic between OVS and
1561 the controller itself.
1563 A: By default, Open vSwitch assumes that OpenFlow controllers are
1564 connected "in-band", that is, that the controllers are actually
1565 part of the network that is being controlled. In in-band mode,
1566 Open vSwitch sets up special "hidden" flows to make sure that
1567 traffic can make it back and forth between OVS and the controllers.
1568 These hidden flows are higher priority than any flows that can be
1569 set up through OpenFlow, and they are not visible through normal
1570 OpenFlow flow table dumps.
1572 Usually, the hidden flows are desirable and helpful, but
1573 occasionally they can cause unexpected behavior. You can view the
1574 full OpenFlow flow table, including hidden flows, on bridge br0
1577 ovs-appctl bridge/dump-flows br0
1579 to help you debug. The hidden flows are those with priorities
1580 greater than 65535 (the maximum priority that can be set with
1583 The DESIGN file at the top level of the Open vSwitch source
1584 distribution describes the in-band model in detail.
1586 If your controllers are not actually in-band (e.g. they are on
1587 localhost via 127.0.0.1, or on a separate network), then you should
1588 configure your controllers in "out-of-band" mode. If you have one
1589 controller on bridge br0, then you can configure out-of-band mode
1592 ovs-vsctl set controller br0 connection-mode=out-of-band
1594 ### Q: I configured all my controllers for out-of-band control mode but
1595 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1597 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1598 set-manager"). By default, Open vSwitch assumes that managers need
1599 in-band rules set up on every bridge. You can disable these rules
1602 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1604 This actually disables in-band control entirely for the bridge, as
1605 if all the bridge's controllers were configured for out-of-band
1608 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1610 A: See answer under "VLANs", above.
1612 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1613 but I got a funny message like this:
1615 ofp_util|INFO|normalization changed ofp_match, details:
1616 ofp_util|INFO| pre: nw_dst=192.168.0.1
1619 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1620 match had disappeared, so that the flow ends up matching every
1623 A: The term "normalization" in the log message means that a flow
1624 cannot match on an L3 field without saying what L3 protocol is in
1625 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1626 so the L3 field match was dropped.
1628 In this case, the L3 protocol could be IP or ARP. A correct
1629 command for each possibility is, respectively:
1631 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1635 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1637 Similarly, a flow cannot match on an L4 field without saying what
1638 L4 protocol is in use. For example, the flow match "tp_src=1234"
1639 is, by itself, meaningless and will be ignored. Instead, to match
1640 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1641 source port 1234, write "udp,tp_src=1234".
1643 ### Q: How can I figure out the OpenFlow port number for a given port?
1645 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1646 respond with an OFPT_FEATURES_REPLY that, among other information,
1647 includes a mapping between OpenFlow port names and numbers. From a
1648 command prompt, "ovs-ofctl show br0" makes such a request and
1649 prints the response for switch br0.
1651 The Interface table in the Open vSwitch database also maps OpenFlow
1652 port names to numbers. To print the OpenFlow port number
1653 associated with interface eth0, run:
1655 ovs-vsctl get Interface eth0 ofport
1657 You can print the entire mapping with:
1659 ovs-vsctl -- --columns=name,ofport list Interface
1661 but the output mixes together interfaces from all bridges in the
1662 database, so it may be confusing if more than one bridge exists.
1664 In the Open vSwitch database, ofport value -1 means that the
1665 interface could not be created due to an error. (The Open vSwitch
1666 log should indicate the reason.) ofport value [] (the empty set)
1667 means that the interface hasn't been created yet. The latter is
1668 normally an intermittent condition (unless ovs-vswitchd is not
1671 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1672 run "ovs-dpctl dump-flows" I don't see them.
1674 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1675 won't display the information that you want. You want to use
1676 "ovs-ofctl dump-flows" instead.
1678 ### Q: It looks like each of the interfaces in my bonded port shows up
1679 as an individual OpenFlow port. Is that right?
1681 A: Yes, Open vSwitch makes individual bond interfaces visible as
1682 OpenFlow ports, rather than the bond as a whole. The interfaces
1683 are treated together as a bond for only a few purposes:
1685 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1686 controller is not configured, this happens implicitly to
1689 - Mirrors configured for output to a bonded port.
1691 It would make a lot of sense for Open vSwitch to present a bond as
1692 a single OpenFlow port. If you want to contribute an
1693 implementation of such a feature, please bring it up on the Open
1694 vSwitch development mailing list at dev@openvswitch.org.
1696 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1697 multiple hosts, and other components. The behavior isn't what I
1700 A: To debug network behavior problems, trace the path of a packet,
1701 hop-by-hop, from its origin in one host to a remote host. If
1702 that's correct, then trace the path of the response packet back to
1705 The open source tool called "plotnetcfg" can help to understand the
1706 relationship between the networking devices on a single host.
1708 Usually a simple ICMP echo request and reply ("ping") packet is
1709 good enough. Start by initiating an ongoing "ping" from the origin
1710 host to a remote host. If you are tracking down a connectivity
1711 problem, the "ping" will not display any successful output, but
1712 packets are still being sent. (In this case the packets being sent
1713 are likely ARP rather than ICMP.)
1715 Tools available for tracing include the following:
1717 - "tcpdump" and "wireshark" for observing hops across network
1718 devices, such as Open vSwitch internal devices and physical
1721 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1722 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1723 These tools allow one to observe the actions being taken on
1724 packets in ongoing flows.
1726 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1727 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1728 documentation, and "Why are there so many different ways to
1729 dump flows?" above for some background.
1731 - "ovs-appctl ofproto/trace" to observe the logic behind how
1732 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1733 documentation. You can out more details about a given flow
1734 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1735 a flow from the output into an "ovs-appctl ofproto/trace"
1738 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1739 observe what goes on at these physical hops.
1741 Starting at the origin of a given packet, observe the packet at
1742 each hop in turn. For example, in one plausible scenario, you
1745 1. "tcpdump" the "eth" interface through which an ARP egresses
1746 a VM, from inside the VM.
1748 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1749 ingresses the host machine.
1751 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1752 the host interface through which the ARP egresses the
1753 physical machine. You may need to use "ovs-dpctl show" to
1754 interpret the port numbers. If the output seems surprising,
1755 you can use "ovs-appctl ofproto/trace" to observe details of
1756 how ovs-vswitchd determined the actions in the "ovs-dpctl
1759 4. "tcpdump" the "eth" interface through which the ARP egresses
1760 the physical machine.
1762 5. "tcpdump" the "eth" interface through which the ARP
1763 ingresses the physical machine, at the remote host that
1766 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1767 remote host that receives the ARP and observe the VM "vif"
1768 or "tap" interface to which the flow is directed. Again,
1769 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1771 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1774 8. "tcpdump" the "eth" interface through which the ARP
1775 ingresses a VM, from inside the VM.
1777 It is likely that during one of these steps you will figure out the
1778 problem. If not, then follow the ARP reply back to the origin, in
1781 ### Q: How do I make a flow drop packets?
1783 A: To drop a packet is to receive it without forwarding it. OpenFlow
1784 explicitly specifies forwarding actions. Thus, a flow with an
1785 empty set of actions does not forward packets anywhere, causing
1786 them to be dropped. You can specify an empty set of actions with
1787 "actions=" on the ovs-ofctl command line. For example:
1789 ovs-ofctl add-flow br0 priority=65535,actions=
1791 would cause every packet entering switch br0 to be dropped.
1793 You can write "drop" explicitly if you like. The effect is the
1794 same. Thus, the following command also causes every packet
1795 entering switch br0 to be dropped:
1797 ovs-ofctl add-flow br0 priority=65535,actions=drop
1799 "drop" is not an action, either in OpenFlow or Open vSwitch.
1800 Rather, it is only a way to say that there are no actions.
1802 ### Q: I added a flow to send packets out the ingress port, like this:
1804 ovs-ofctl add-flow br0 in_port=2,actions=2
1806 but OVS drops the packets instead.
1808 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1809 out its ingress port. The rationale is that dropping these packets
1810 makes it harder to loop the network. Sometimes this behavior can
1811 even be convenient, e.g. it is often the desired behavior in a flow
1812 that forwards a packet to several ports ("floods" the packet).
1814 Sometimes one really needs to send a packet out its ingress port
1815 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1816 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1818 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1820 This also works in some circumstances where the flow doesn't match
1821 on the input port. For example, if you know that your switch has
1822 five ports numbered 2 through 6, then the following will send every
1823 received packet out every port, even its ingress port:
1825 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1829 ovs-ofctl add-flow br0 actions=all,in_port
1831 Sometimes, in complicated flow tables with multiple levels of
1832 "resubmit" actions, a flow needs to output to a particular port
1833 that may or may not be the ingress port. It's difficult to take
1834 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1835 provides, as an OpenFlow extension, the ability to modify the
1836 in_port field. Whatever value is currently in the in_port field is
1837 the port to which outputs will be dropped, as well as the
1838 destination for OFPP_IN_PORT. This means that the following will
1839 reliably output to port 2 or to ports 2 through 6, respectively:
1841 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1842 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1844 If the input port is important, then one may save and restore it on
1847 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1848 load:0->NXM_OF_IN_PORT[],\
1850 pop:NXM_OF_IN_PORT[]
1852 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1853 on port 2. I set up flows to forward only traffic destined to the
1854 other host and drop other traffic, like this:
1856 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1857 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1858 priority=0,actions=drop
1860 But it doesn't work--I don't get any connectivity when I do this.
1863 A: These flows drop the ARP packets that IP hosts use to establish IP
1864 connectivity over Ethernet. To solve the problem, add flows to
1865 allow ARP to pass between the hosts:
1867 priority=5,in_port=1,arp,actions=2
1868 priority=5,in_port=2,arp,actions=1
1870 This issue can manifest other ways, too. The following flows that
1871 match on Ethernet addresses instead of IP addresses will also drop
1872 ARP packets, because ARP requests are broadcast instead of being
1873 directed to a specific host:
1875 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1876 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1877 priority=0,actions=drop
1879 The solution already described above will also work in this case.
1880 It may be better to add flows to allow all multicast and broadcast
1883 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1884 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1886 ### Q: My bridge disconnects from my controller on add-port/del-port.
1888 A: Reconfiguring your bridge can change your bridge's datapath-id because
1889 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1890 In that case, Open vSwitch disconnects from controllers because there's
1891 no graceful way to notify controllers about the change of datapath-id.
1893 To avoid the behaviour, you can configure datapath-id manually.
1895 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1897 ### Q: My controller is getting errors about "buffers". What's going on?
1899 A: When a switch sends a packet to an OpenFlow controller using a
1900 "packet-in" message, it can also keep a copy of that packet in a
1901 "buffer", identified by a 32-bit integer "buffer_id". There are
1902 two advantages to buffering. First, when the controller wants to
1903 tell the switch to do something with the buffered packet (with a
1904 "packet-out" OpenFlow request), it does not need to send another
1905 copy of the packet back across the OpenFlow connection, which
1906 reduces the bandwidth cost of the connection and improves latency.
1907 This enables the second advantage: the switch can optionally send
1908 only the first part of the packet to the controller (assuming that
1909 the switch only needs to look at the first few bytes of the
1910 packet), further reducing bandwidth and improving latency.
1912 However, buffering introduces some issues of its own. First, any
1913 switch has limited resources, so if the controller does not use a
1914 buffered packet, the switch has to decide how long to keep it
1915 buffered. When many packets are sent to a controller and buffered,
1916 Open vSwitch can discard buffered packets that the controller has
1917 not used after as little as 5 seconds. This means that
1918 controllers, if they make use of packet buffering, should use the
1919 buffered packets promptly. (This includes sending a "packet-out"
1920 with no actions if the controller does not want to do anything with
1921 a buffered packet, to clear the packet buffer and effectively
1924 Second, packet buffers are one-time-use, meaning that a controller
1925 cannot use a single packet buffer in two or more "packet-out"
1926 commands. Open vSwitch will respond with an error to the second
1927 and subsequent "packet-out"s in such a case.
1929 Finally, a common error early in controller development is to try
1930 to use buffer_id 0 in a "packet-out" message as if 0 represented
1931 "no buffered packet". This is incorrect usage: the buffer_id with
1932 this meaning is actually 0xffffffff.
1934 ovs-vswitchd(8) describes some details of Open vSwitch packet
1935 buffering that the OpenFlow specification requires implementations
1942 ### Q: How do I implement a new OpenFlow message?
1944 A: Add your new message to "enum ofpraw" and "enum ofptype" in
1945 lib/ofp-msgs.h, following the existing pattern. Then recompile and
1946 fix all of the new warnings, implementing new functionality for the
1947 new message as needed. (If you configure with --enable-Werror, as
1948 described in [INSTALL.md], then it is impossible to miss any warnings.)
1950 If you need to add an OpenFlow vendor extension message for a
1951 vendor that doesn't yet have any extension messages, then you will
1952 also need to edit build-aux/extract-ofp-msgs.
1954 ### Q: How do I add support for a new field or header?
1956 A: Add new members for your field to "struct flow" in lib/flow.h, and
1957 add new enumerations for your new field to "enum mf_field_id" in
1958 lib/meta-flow.h, following the existing pattern. Also, add support
1959 to miniflow_extract() in lib/flow.c for extracting your new field
1960 from a packet into struct miniflow. Then recompile and fix all of
1961 the new warnings, implementing new functionality for the new field
1962 or header as needed. (If you configure with --enable-Werror, as
1963 described in [INSTALL.md], then it is impossible to miss any
1966 If you want kernel datapath support for your new field, you also
1967 need to modify the kernel module for the operating systems you are
1968 interested in. This isn't mandatory, since fields understood only
1969 by userspace work too (with a performance penalty), so it's
1970 reasonable to start development without it. If you implement
1971 kernel module support for Linux, then the Linux kernel "netdev"
1972 mailing list is the place to submit that support first; please read
1973 up on the Linux kernel development process separately. The Windows
1974 datapath kernel module support, on the other hand, is maintained
1975 within the OVS tree, so patches for that can go directly to
1978 ### Q: How do I add support for a new OpenFlow action?
1980 A: Add your new action to "enum ofp_raw_action_type" in
1981 lib/ofp-actions.c, following the existing pattern. Then recompile
1982 and fix all of the new warnings, implementing new functionality for
1983 the new action as needed. (If you configure with --enable-Werror,
1984 as described in [INSTALL.md], then it is impossible to miss any
1987 If you need to add an OpenFlow vendor extension action for a vendor
1988 that doesn't yet have any extension actions, then you will also
1989 need to edit build-aux/extract-ofp-actions.
1995 bugs@openvswitch.org
1996 http://openvswitch.org/
1998 [PORTING.md]:PORTING.md
1999 [WHY-OVS.md]:WHY-OVS.md
2000 [INSTALL.md]:INSTALL.md
2001 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
2002 [INSTALL.DPDK.md]:INSTALL.DPDK.md
projects / cascardo / ovs.git / blob