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
160 Open vSwitch userspace should also work with the Linux kernel module
161 built into Linux 3.3 and later.
163 Open vSwitch userspace is not sensitive to the Linux kernel version.
164 It should build against almost any kernel, certainly against 2.6.32
167 ### Q: I get an error like this when I configure Open vSwitch:
169 configure: error: Linux kernel in <dir> is version <x>, but
170 version newer than <y> is not supported (please refer to the
175 A: You have the following options:
177 - Use the Linux kernel module supplied with the kernel that you are
178 using. (See also the following FAQ.)
180 - If there is a newer released version of Open vSwitch, consider
181 building that one, because it may support the kernel that you are
182 building against. (To find out, consult the table in the
185 - The Open vSwitch "master" branch may support the kernel that you
186 are using, so consider building the kernel module from "master".
188 All versions of Open vSwitch userspace are compatible with all
189 versions of the Open vSwitch kernel module, so you do not have to
190 use the kernel module from one source along with the userspace
191 programs from the same source.
193 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
195 A: The kernel module in upstream Linux does not include support for
196 LISP. Work is in progress to add support for LISP to the upstream
197 Linux version of the Open vSwitch kernel module. For now, if you
198 need this feature, use the kernel module from the Open vSwitch
199 distribution instead of the upstream Linux kernel module.
201 Certain features require kernel support to function or to have
202 reasonable performance. If the ovs-vswitchd log file indicates that
203 a feature is not supported, consider upgrading to a newer upstream
204 Linux release or using the kernel module paired with the userspace
207 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
209 A: Support for tunnels was added to the upstream Linux kernel module
210 after the rest of Open vSwitch. As a result, some kernels may contain
211 support for Open vSwitch but not tunnels. The minimum kernel version
212 that supports each tunnel protocol is:
214 | Protocol | Linux Kernel
215 |:--------:|:-------------:
219 | LISP | <not upstream>
220 | STT | <not upstream>
222 If you are using a version of the kernel that is older than the one
223 listed above, it is still possible to use that tunnel protocol. However,
224 you must compile and install the kernel module included with the Open
225 vSwitch distribution rather than the one on your machine. If problems
226 persist after doing this, check to make sure that the module that is
227 loaded is the one you expect.
229 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
231 A: Generating outer UDP checksums requires kernel support that was not
232 part of the initial implementation of these protocols. If using the
233 upstream Linux Open vSwitch module, you must use kernel 4.0 or
234 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
235 support UDP checksums.
237 ### Q: What features are not available when using the userspace datapath?
239 A: Tunnel virtual ports are not supported, as described in the
240 previous answer. It is also not possible to use queue-related
241 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
242 may not be transmitted.
244 ### Q: What Linux kernel versions does IPFIX flow monitoring work with?
246 A: IPFIX flow monitoring requires the Linux kernel module from Linux
247 3.10 or later, or the out-of-tree module from Open vSwitch version
250 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
252 In general, the Open vSwitch userspace should be used with the
253 kernel version included in the same release or with the version
254 from upstream Linux. However, when upgrading between two releases
255 of Open vSwitch it is best to migrate userspace first to reduce
256 the possibility of incompatibilities.
258 ### Q: What happened to the bridge compatibility feature?
260 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
261 When it was enabled, Open vSwitch imitated the interface of the
262 Linux kernel "bridge" module. This allowed users to drop Open
263 vSwitch into environments designed to use the Linux kernel bridge
264 module without adapting the environment to use Open vSwitch.
266 Open vSwitch 1.10 and later do not support bridge compatibility.
267 The feature was dropped because version 1.10 adopted a new internal
268 architecture that made bridge compatibility difficult to maintain.
269 Now that many environments use OVS directly, it would be rarely
272 To use bridge compatibility, install OVS 1.9 or earlier, including
273 the accompanying kernel modules (both the main and bridge
274 compatibility modules), following the instructions that come with
275 the release. Be sure to start the ovs-brcompatd daemon.
281 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
283 A: In networking, the terms "bridge" and "switch" are synonyms. Open
284 vSwitch implements an Ethernet switch, which means that it is also
287 ### Q: What's a VLAN?
289 A: See the "VLAN" section below.
295 ### Q: How do I configure a port as an access port?
297 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
298 the following commands configure br0 with eth0 as a trunk port (the
299 default) and tap0 as an access port for VLAN 9:
302 ovs-vsctl add-port br0 eth0
303 ovs-vsctl add-port br0 tap0 tag=9
305 If you want to configure an already added port as an access port,
306 use "ovs-vsctl set", e.g.:
308 ovs-vsctl set port tap0 tag=9
310 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
312 A: The following commands configure br0 with eth0 and tap0 as trunk
313 ports. All traffic coming in or going out on eth0 or tap0 is also
314 mirrored to tap1; any traffic arriving on tap1 is dropped:
317 ovs-vsctl add-port br0 eth0
318 ovs-vsctl add-port br0 tap0
319 ovs-vsctl add-port br0 tap1 \
320 -- --id=@p get port tap1 \
321 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
322 -- set bridge br0 mirrors=@m
324 To later disable mirroring, run:
326 ovs-vsctl clear bridge br0 mirrors
328 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
330 A: Yes. How you configure it depends on what you mean by "promiscuous
333 - Conventionally, "promiscuous mode" is a feature of a network
334 interface card. Ordinarily, a NIC passes to the CPU only the
335 packets actually destined to its host machine. It discards
336 the rest to avoid wasting memory and CPU cycles. When
337 promiscuous mode is enabled, however, it passes every packet
338 to the CPU. On an old-style shared-media or hub-based
339 network, this allows the host to spy on all packets on the
340 network. But in the switched networks that are almost
341 everywhere these days, promiscuous mode doesn't have much
342 effect, because few packets not destined to a host are
343 delivered to the host's NIC.
345 This form of promiscuous mode is configured in the guest OS of
346 the VMs on your bridge, e.g. with "ifconfig".
348 - The VMware vSwitch uses a different definition of "promiscuous
349 mode". When you configure promiscuous mode on a VMware vNIC,
350 the vSwitch sends a copy of every packet received by the
351 vSwitch to that vNIC. That has a much bigger effect than just
352 enabling promiscuous mode in a guest OS. Rather than getting
353 a few stray packets for which the switch does not yet know the
354 correct destination, the vNIC gets every packet. The effect
355 is similar to replacing the vSwitch by a virtual hub.
357 This "promiscuous mode" is what switches normally call "port
358 mirroring" or "SPAN". For information on how to configure
359 SPAN, see "How do I configure a port as a SPAN port, that is,
360 enable mirroring of all traffic to that port?"
362 ### Q: How do I configure a DPDK port as an access port?
364 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
366 If your version is DPDK-enabled it will support the --dpdk
367 argument on the command line and will display lines with
368 "EAL:..." during startup when --dpdk is supplied.
370 Secondly, when adding a DPDK port, unlike a system port, the
371 type for the interface must be specified. For example;
374 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
376 Finally, it is required that DPDK port names begin with 'dpdk'.
378 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
381 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
383 A: The following commands configure br0 with eth0 as a trunk port and
384 tap0 as an access port for VLAN 10. All traffic coming in or going
385 out on tap0, as well as traffic coming in or going out on eth0 in
386 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
387 VLAN 10, in cases where one is present, is dropped as part of
391 ovs-vsctl add-port br0 eth0
392 ovs-vsctl add-port br0 tap0 tag=10
394 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
396 -- set bridge br0 mirrors=@m
398 To later disable mirroring, run:
400 ovs-vsctl clear bridge br0 mirrors
402 Mirroring to a VLAN can disrupt a network that contains unmanaged
403 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
404 GRE tunnel has fewer caveats than mirroring to a VLAN and should
405 generally be preferred.
407 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
409 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
410 of the specified output-vlan. This loss of information may make
411 the mirrored traffic too hard to interpret.
413 To mirror multiple VLANs, use the commands above, but specify a
414 comma-separated list of VLANs as the value for select-vlan. To
415 mirror every VLAN, use the commands above, but omit select-vlan and
418 When a packet arrives on a VLAN that is used as a mirror output
419 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
420 floods the packet across all the ports for which the mirror output
421 VLAN is configured. (If an OpenFlow controller is in use, then it
422 can override this behavior through the flow table.) If OVS is used
423 as an intermediate switch, rather than an edge switch, this ensures
424 that the RSPAN traffic is distributed through the network.
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: How do I configure mirroring of all traffic to a GRE tunnel?
433 A: The following commands configure br0 with eth0 and tap0 as trunk
434 ports. All traffic coming in or going out on eth0 or tap0 is also
435 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
436 traffic arriving on gre0 is dropped:
439 ovs-vsctl add-port br0 eth0
440 ovs-vsctl add-port br0 tap0
441 ovs-vsctl add-port br0 gre0 \
442 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
443 -- --id=@p get port gre0 \
444 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
445 -- set bridge br0 mirrors=@m
447 To later disable mirroring and destroy the GRE tunnel:
449 ovs-vsctl clear bridge br0 mirrors
450 ovs-vcstl del-port br0 gre0
452 ### Q: Does Open vSwitch support ERSPAN?
454 A: No. ERSPAN is an undocumented proprietary protocol. As an
455 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
458 ### Q: How do I connect two bridges?
460 A: First, why do you want to do this? Two connected bridges are not
461 much different from a single bridge, so you might as well just have
462 a single bridge with all your ports on it.
464 If you still want to connect two bridges, you can use a pair of
465 patch ports. The following example creates bridges br0 and br1,
466 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
467 and br1 with a pair of patch ports.
470 ovs-vsctl add-port br0 eth0
471 ovs-vsctl add-port br0 tap0
473 ovs-vsctl add-port br1 tap1
475 -- add-port br0 patch0 \
476 -- set interface patch0 type=patch options:peer=patch1 \
477 -- add-port br1 patch1 \
478 -- set interface patch1 type=patch options:peer=patch0
480 Bridges connected with patch ports are much like a single bridge.
481 For instance, if the example above also added eth1 to br1, and both
482 eth0 and eth1 happened to be connected to the same next-hop switch,
483 then you could loop your network just as you would if you added
484 eth0 and eth1 to the same bridge (see the "Configuration Problems"
485 section below for more information).
487 If you are using Open vSwitch 1.9 or an earlier version, then you
488 need to be using the kernel module bundled with Open vSwitch rather
489 than the one that is integrated into Linux 3.3 and later, because
490 Open vSwitch 1.9 and earlier versions need kernel support for patch
491 ports. This also means that in Open vSwitch 1.9 and earlier, patch
492 ports will not work with the userspace datapath, only with the
495 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
497 A: Open vSwitch does not support such a configuration.
498 Bridges always have their local ports.
501 Implementation Details
502 ----------------------
504 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
506 A: Open vSwitch uses different kinds of flows for different purposes:
508 - OpenFlow flows are the most important kind of flow. OpenFlow
509 controllers use these flows to define a switch's policy.
510 OpenFlow flows support wildcards, priorities, and multiple
513 When in-band control is in use, Open vSwitch sets up a few
514 "hidden" flows, with priority higher than a controller or the
515 user can configure, that are not visible via OpenFlow. (See
516 the "Controller" section of the FAQ for more information
519 - The Open vSwitch software switch implementation uses a second
520 kind of flow internally. These flows, called "datapath" or
521 "kernel" flows, do not support priorities and comprise only a
522 single table, which makes them suitable for caching. (Like
523 OpenFlow flows, datapath flows do support wildcarding, in Open
524 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
525 also support different actions and number ports differently.
527 Datapath flows are an implementation detail that is subject to
528 change in future versions of Open vSwitch. Even with the
529 current version of Open vSwitch, hardware switch
530 implementations do not necessarily use this architecture.
532 Users and controllers directly control only the OpenFlow flow
533 table. Open vSwitch manages the datapath flow table itself, so
534 users should not normally be concerned with it.
536 ### Q: Why are there so many different ways to dump flows?
538 A: Open vSwitch has two kinds of flows (see the previous question), so
539 it has commands with different purposes for dumping each kind of
542 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
543 hidden flows. This is the most commonly useful form of flow
544 dump. (Unlike the other commands, this should work with any
545 OpenFlow switch, not just Open vSwitch.)
547 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
548 including hidden flows. This is occasionally useful for
549 troubleshooting suspected issues with in-band control.
551 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
552 entries for a Linux kernel-based datapath. In Open vSwitch
553 1.10 and later, ovs-vswitchd merges multiple switches into a
554 single datapath, so it will show all the flows on all your
555 kernel-based switches. This command can occasionally be
556 useful for debugging.
558 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
559 dumps datapath flows for only the specified bridge, regardless
562 ### Q: How does multicast snooping works with VLANs?
564 A: Open vSwitch maintains snooping tables for each VLAN.
570 ### Q: I just upgraded and I see a performance drop. Why?
572 A: The OVS kernel datapath may have been updated to a newer version than
573 the OVS userspace components. Sometimes new versions of OVS kernel
574 module add functionality that is backwards compatible with older
575 userspace components but may cause a drop in performance with them.
576 Especially, if a kernel module from OVS 2.1 or newer is paired with
577 OVS userspace 1.10 or older, there will be a performance drop for
580 Updating the OVS userspace components to the latest released
581 version should fix the performance degradation.
583 To get the best possible performance and functionality, it is
584 recommended to pair the same versions of the kernel module and OVS
588 Configuration Problems
589 ----------------------
591 ### Q: I created a bridge and added my Ethernet port to it, using commands
595 ovs-vsctl add-port br0 eth0
597 and as soon as I ran the "add-port" command I lost all connectivity
600 A: A physical Ethernet device that is part of an Open vSwitch bridge
601 should not have an IP address. If one does, then that IP address
602 will not be fully functional.
604 You can restore functionality by moving the IP address to an Open
605 vSwitch "internal" device, such as the network device named after
606 the bridge itself. For example, assuming that eth0's IP address is
607 192.168.128.5, you could run the commands below to fix up the
610 ifconfig eth0 0.0.0.0
611 ifconfig br0 192.168.128.5
613 (If your only connection to the machine running OVS is through the
614 IP address in question, then you would want to run all of these
615 commands on a single command line, or put them into a script.) If
616 there were any additional routes assigned to eth0, then you would
617 also want to use commands to adjust these routes to go through br0.
619 If you use DHCP to obtain an IP address, then you should kill the
620 DHCP client that was listening on the physical Ethernet interface
621 (e.g. eth0) and start one listening on the internal interface
622 (e.g. br0). You might still need to manually clear the IP address
623 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
625 There is no compelling reason why Open vSwitch must work this way.
626 However, this is the way that the Linux kernel bridge module has
627 always worked, so it's a model that those accustomed to Linux
628 bridging are already used to. Also, the model that most people
629 expect is not implementable without kernel changes on all the
630 versions of Linux that Open vSwitch supports.
632 By the way, this issue is not specific to physical Ethernet
633 devices. It applies to all network devices except Open vSwitch
636 ### Q: I created a bridge and added a couple of Ethernet ports to it,
637 ### using commands like these:
640 ovs-vsctl add-port br0 eth0
641 ovs-vsctl add-port br0 eth1
643 and now my network seems to have melted: connectivity is unreliable
644 (even connectivity that doesn't go through Open vSwitch), all the
645 LEDs on my physical switches are blinking, wireshark shows
646 duplicated packets, and CPU usage is very high.
648 A: More than likely, you've looped your network. Probably, eth0 and
649 eth1 are connected to the same physical Ethernet switch. This
650 yields a scenario where OVS receives a broadcast packet on eth0 and
651 sends it out on eth1, then the physical switch connected to eth1
652 sends the packet back on eth0, and so on forever. More complicated
653 scenarios, involving a loop through multiple switches, are possible
656 The solution depends on what you are trying to do:
658 - If you added eth0 and eth1 to get higher bandwidth or higher
659 reliability between OVS and your physical Ethernet switch,
660 use a bond. The following commands create br0 and then add
661 eth0 and eth1 as a bond:
664 ovs-vsctl add-bond br0 bond0 eth0 eth1
666 Bonds have tons of configuration options. Please read the
667 documentation on the Port table in ovs-vswitchd.conf.db(5)
670 Configuration for DPDK-enabled interfaces is slightly less
671 straightforward: see [INSTALL.DPDK.md].
673 - Perhaps you don't actually need eth0 and eth1 to be on the
674 same bridge. For example, if you simply want to be able to
675 connect each of them to virtual machines, then you can put
676 each of them on a bridge of its own:
679 ovs-vsctl add-port br0 eth0
682 ovs-vsctl add-port br1 eth1
684 and then connect VMs to br0 and br1. (A potential
685 disadvantage is that traffic cannot directly pass between br0
686 and br1. Instead, it will go out eth0 and come back in eth1,
689 - If you have a redundant or complex network topology and you
690 want to prevent loops, turn on spanning tree protocol (STP).
691 The following commands create br0, enable STP, and add eth0
692 and eth1 to the bridge. The order is important because you
693 don't want have to have a loop in your network even
697 ovs-vsctl set bridge br0 stp_enable=true
698 ovs-vsctl add-port br0 eth0
699 ovs-vsctl add-port br0 eth1
701 The Open vSwitch implementation of STP is not well tested.
702 Please report any bugs you observe, but if you'd rather avoid
703 acting as a beta tester then another option might be your
706 ### Q: I can't seem to use Open vSwitch in a wireless network.
708 A: Wireless base stations generally only allow packets with the source
709 MAC address of NIC that completed the initial handshake.
710 Therefore, without MAC rewriting, only a single device can
711 communicate over a single wireless link.
713 This isn't specific to Open vSwitch, it's enforced by the access
714 point, so the same problems will show up with the Linux bridge or
715 any other way to do bridging.
717 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
719 A: PPP most commonly carries IP packets, but Open vSwitch works only
720 with Ethernet frames. The correct way to interface PPP to an
721 Ethernet network is usually to use routing instead of switching.
723 ### Q: Is there any documentation on the database tables and fields?
725 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
727 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
728 I only see a datapath called "ovs-system". How can I see datapath
729 information about a particular bridge?
731 A: In version 1.9.0, OVS switched to using a single datapath that is
732 shared by all bridges of that type. The "ovs-appctl dpif/*"
733 commands provide similar functionality that is scoped by the bridge.
735 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
736 see the port in the datapath?
738 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
739 cannot be loaded at the same time. It is likely that on your system the
740 Linux GRE module is already loaded and blocking OVS (to confirm, check
741 dmesg for errors regarding GRE registration). To fix this, unload all
742 GRE modules that appear in lsmod as well as the OVS kernel module. You
743 can then reload the OVS module following the directions in
744 [INSTALL.md], which will ensure that dependencies are satisfied.
746 ### Q: Open vSwitch does not seem to obey my packet filter rules.
748 A: It depends on mechanisms and configurations you want to use.
750 You cannot usefully use typical packet filters, like iptables, on
751 physical Ethernet ports that you add to an Open vSwitch bridge.
752 This is because Open vSwitch captures packets from the interface at
753 a layer lower below where typical packet-filter implementations
754 install their hooks. (This actually applies to any interface of
755 type "system" that you might add to an Open vSwitch bridge.)
757 You can usefully use typical packet filters on Open vSwitch
758 internal ports as they are mostly ordinary interfaces from the point
759 of view of packet filters.
761 For example, suppose you create a bridge br0 and add Ethernet port
762 eth0 to it. Then you can usefully add iptables rules to affect the
763 internal interface br0, but not the physical interface eth0. (br0
764 is also where you would add an IP address, as discussed elsewhere
767 For simple filtering rules, it might be possible to achieve similar
768 results by installing appropriate OpenFlow flows instead.
770 If the use of a particular packet filter setup is essential, Open
771 vSwitch might not be the best choice for you. On Linux, you might
772 want to consider using the Linux Bridge. (This is the only choice if
773 you want to use ebtables rules.) On NetBSD, you might want to
774 consider using the bridge(4) with BRIDGE_IPF option.
776 ### Q: It seems that Open vSwitch does nothing when I removed a port and
777 then immediately put it back. For example, consider that p1 is
778 a port of type=internal:
780 ovs-vsctl del-port br0 p1 -- \
782 set interface p1 type=internal
784 A: It's an expected behaviour.
786 If del-port and add-port happen in a single OVSDB transaction as
787 your example, Open vSwitch always "skips" the intermediate steps.
788 Even if they are done in multiple transactions, it's still allowed
789 for Open vSwitch to skip the intermediate steps and just implement
790 the overall effect. In both cases, your example would be turned
793 If you want to make Open vSwitch actually destroy and then re-create
794 the port for some side effects like resetting kernel setting for the
795 corresponding interface, you need to separate operations into multiple
796 OVSDB transactions and ensure that at least the first one does not have
797 --no-wait. In the following example, the first ovs-vsctl will block
798 until Open vSwitch reloads the new configuration and removes the port:
800 ovs-vsctl del-port br0 p1
801 ovs-vsctl add-port br0 p1 -- \
802 set interface p1 type=internal
804 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
805 it takes too long (minutes or hours) to do it with ovs-vsctl. How
808 A: If you add them one at a time with ovs-vsctl, it can take a long
809 time to add thousands of ports to an Open vSwitch bridge. This is
810 because every invocation of ovs-vsctl first reads the current
811 configuration from OVSDB. As the number of ports grows, this
812 starts to take an appreciable amount of time, and when it is
813 repeated thousands of times the total time becomes significant.
815 The solution is to add the ports in one invocation of ovs-vsctl (or
816 a small number of them). For example, using bash:
819 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
822 takes seconds, not minutes or hours, in the OVS sandbox environment.
824 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
825 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
828 A: Open vSwitch wasn't able to create the bridge. Check the
829 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
830 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
832 In general, the Open vSwitch database reflects the desired
833 configuration state. ovs-vswitchd monitors the database and, when
834 it changes, reconfigures the system to reflect the new desired
835 state. This normally happens very quickly. Thus, a discrepancy
836 between the database and the actual state indicates that
837 ovs-vswitchd could not implement the configuration, and so one
838 should check the log to find out why. (Another possible cause is
839 that ovs-vswitchd is not running. This will make "ovs-vsctl"
840 commands hang, if they change the configuration, unless one
841 specifies "--no-wait".)
843 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
844 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
845 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
848 A: Open vSwitch wasn't able to create the port. Check the
849 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
850 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
851 see the previous question for more information.
853 You may want to upgrade to Open vSwitch 2.3 (or later), in which
854 ovs-vsctl will immediately report when there is an issue creating a
857 ### Q: I created a tap device tap0, configured an IP address on it, and
858 added it to a bridge, like this:
861 ifconfig tap0 192.168.0.123
863 ovs-vsctl add-port br0 tap0
865 I expected that I could then use this IP address to contact other
866 hosts on the network, but it doesn't work. Why not?
868 A: The short answer is that this is a misuse of a "tap" device. Use
869 an "internal" device implemented by Open vSwitch, which works
870 differently and is designed for this use. To solve this problem
871 with an internal device, instead run:
874 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
875 ifconfig int0 192.168.0.123
877 Even more simply, you can take advantage of the internal port that
878 every bridge has under the name of the bridge:
881 ifconfig br0 192.168.0.123
883 In more detail, a "tap" device is an interface between the Linux
884 (or *BSD) network stack and a user program that opens it as a
885 socket. When the "tap" device transmits a packet, it appears in
886 the socket opened by the userspace program. Conversely, when the
887 userspace program writes to the "tap" socket, the kernel TCP/IP
888 stack processes the packet as if it had been received by the "tap"
891 Consider the configuration above. Given this configuration, if you
892 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
893 routing stack will transmit an ARP on the tap0 device. Open
894 vSwitch userspace treats "tap" devices just like any other network
895 device; that is, it doesn't open them as "tap" sockets. That means
896 that the ARP packet will simply get dropped.
898 You might wonder why the Open vSwitch kernel module doesn't
899 intercept the ARP packet and bridge it. After all, Open vSwitch
900 intercepts packets on other devices. The answer is that Open
901 vSwitch only intercepts *received* packets, but this is a packet
902 being transmitted. The same thing happens for all other types of
903 network devices, except for Open vSwitch "internal" ports. If you,
904 for example, add a physical Ethernet port to an OVS bridge,
905 configure an IP address on a physical Ethernet port, and then issue
906 a "ping" to an address in that subnet, the same thing happens: an
907 ARP gets transmitted on the physical Ethernet port and Open vSwitch
908 never sees it. (You should not do that, as documented at the
909 beginning of this section.)
911 It can make sense to add a "tap" device to an Open vSwitch bridge,
912 if some userspace program (other than Open vSwitch) has opened the
913 tap socket. This is the case, for example, if the "tap" device was
914 created by KVM (or QEMU) to simulate a virtual NIC. In such a
915 case, when OVS bridges a packet to the "tap" device, the kernel
916 forwards that packet to KVM in userspace, which passes it along to
917 the VM, and in the other direction, when the VM sends a packet, KVM
918 writes it to the "tap" socket, which causes OVS to receive it and
919 bridge it to the other OVS ports. Please note that in such a case
920 no IP address is configured on the "tap" device (there is normally
921 an IP address configured in the virtual NIC inside the VM, but this
922 is not visible to the host Linux kernel or to Open vSwitch).
924 There is one special case in which Open vSwitch does directly read
925 and write "tap" sockets. This is an implementation detail of the
926 Open vSwitch userspace switch, which implements its "internal"
927 ports as Linux (or *BSD) "tap" sockets. In such a userspace
928 switch, OVS receives packets sent on the "tap" device used to
929 implement an "internal" port by reading the associated "tap"
930 socket, and bridges them to the rest of the switch. In the other
931 direction, OVS transmits packets bridged to the "internal" port by
932 writing them to the "tap" socket, causing them to be processed by
933 the kernel TCP/IP stack as if they had been received on the "tap"
934 device. Users should not need to be concerned with this
935 implementation detail.
937 Open vSwitch has a network device type called "tap". This is
938 intended only for implementing "internal" ports in the OVS
939 userspace switch and should not be used otherwise. In particular,
940 users should not configure KVM "tap" devices as type "tap" (use
941 type "system", the default, instead).
944 Quality of Service (QoS)
945 ------------------------
947 ### Q: Does OVS support Quality of Service (QoS)?
949 A: Yes. For traffic that egresses from a switch, OVS supports traffic
950 shaping; for traffic that ingresses into a switch, OVS support
951 policing. Policing is a simple form of quality-of-service that
952 simply drops packets received in excess of the configured rate. Due
953 to its simplicity, policing is usually less accurate and less
954 effective than egress traffic shaping, which queues packets.
956 Keep in mind that ingress and egress are from the perspective of the
957 switch. That means that egress shaping limits the rate at which
958 traffic is allowed to transmit from a physical interface, but the
959 rate at which traffic will be received on a virtual machine's VIF.
960 For ingress policing, the behavior is the opposite.
962 ### Q: How do I configure egress traffic shaping?
964 A: Suppose that you want to set up bridge br0 connected to physical
965 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
966 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
967 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
968 could configure the bridge this way:
972 add-port br0 eth0 -- \
973 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
974 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
975 set port eth0 qos=@newqos -- \
976 --id=@newqos create qos type=linux-htb \
977 other-config:max-rate=1000000000 \
978 queues:123=@vif10queue \
979 queues:234=@vif20queue -- \
980 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
981 --id=@vif20queue create queue other-config:max-rate=20000000
983 At this point, bridge br0 is configured with the ports and eth0 is
984 configured with the queues that you need for QoS, but nothing is
985 actually directing packets from vif1.0 or vif2.0 to the queues that
986 we have set up for them. That means that all of the packets to
987 eth0 are going to the "default queue", which is not what we want.
989 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
990 queues reserved for them:
992 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
993 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
995 Each of the above flows matches on the input port, sets up the
996 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
997 executes the "normal" action, which performs the same switching
998 that Open vSwitch would have done without any OpenFlow flows being
999 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1000 numbers 5 and 6, respectively, because we set their ofport_request
1001 columns above. If we had not done that, then we would have needed
1002 to find out their port numbers before setting up these flows.)
1004 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1007 By the way, if you delete the bridge created by the above commands,
1010 ovs-vsctl del-br br0
1012 then that will leave one unreferenced QoS record and two
1013 unreferenced Queue records in the Open vSwich database. One way to
1014 clear them out, assuming you don't have other QoS or Queue records
1015 that you want to keep, is:
1017 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1019 If you do want to keep some QoS or Queue records, or the Open
1020 vSwitch you are using is older than version 1.8 (which added the
1021 --all option), then you will have to destroy QoS and Queue records
1024 ### Q: How do I configure ingress policing?
1026 A: A policing policy can be configured on an interface to drop packets
1027 that arrive at a higher rate than the configured value. For example,
1028 the following commands will rate-limit traffic that vif1.0 may
1031 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1032 ovs-vsctl set interface vif1.0 ingress_policing_burst=1000
1034 Traffic policing can interact poorly with some network protocols and
1035 can have surprising results. The "Ingress Policing" section of
1036 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1038 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1039 adding records to the QoS and Queue table, but the results aren't
1042 A: Did you install OpenFlow flows that use your queues? This is the
1043 primary way to tell Open vSwitch which queues you want to use. If
1044 you don't do this, then the default queue will be used, which will
1045 probably not have the effect you want.
1047 Refer to the previous question for an example.
1049 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1050 doesn't yet support. How do I do that?
1052 A: Open vSwitch does not implement QoS itself. Instead, it can
1053 configure some, but not all, of the QoS features built into the
1054 Linux kernel. If you need some QoS feature that OVS cannot
1055 configure itself, then the first step is to figure out whether
1056 Linux QoS supports that feature. If it does, then you can submit a
1057 patch to support Open vSwitch configuration for that feature, or
1058 you can use "tc" directly to configure the feature in Linux. (If
1059 Linux QoS doesn't support the feature you want, then first you have
1060 to add that support to Linux.)
1062 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1063 working as well as I expect.
1065 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1068 - Open vSwitch configures a subset of Linux kernel QoS
1069 features, according to what is in OVSDB. It is possible that
1070 this code has bugs. If you believe that this is so, then you
1071 can configure the Linux traffic control (QoS) stack directly
1072 with the "tc" program. If you get better results that way,
1073 you can send a detailed bug report to bugs@openvswitch.org.
1075 It is certain that Open vSwitch cannot configure every Linux
1076 kernel QoS feature. If you need some feature that OVS cannot
1077 configure, then you can also use "tc" directly (or add that
1080 - The Open vSwitch implementation of OpenFlow allows flows to
1081 be directed to particular queues. This is pretty simple and
1082 unlikely to have serious bugs at this point.
1084 However, most problems with QoS on Linux are not bugs in Open
1085 vSwitch at all. They tend to be either configuration errors
1086 (please see the earlier questions in this section) or issues with
1087 the traffic control (QoS) stack in Linux. The Open vSwitch
1088 developers are not experts on Linux traffic control. We suggest
1089 that, if you believe you are encountering a problem with Linux
1090 traffic control, that you consult the tc manpages (e.g. tc(8),
1091 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1092 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1094 ### Q: Does Open vSwitch support OpenFlow meters?
1096 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1097 OpenFlow meters. There is no implementation of meters in the Open
1098 vSwitch software switch (neither the kernel-based nor userspace
1105 ### Q: What's a VLAN?
1107 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1108 partition a single switch into multiple switches. Suppose, for
1109 example, that you have two groups of machines, group A and group B.
1110 You want the machines in group A to be able to talk to each other,
1111 and you want the machine in group B to be able to talk to each
1112 other, but you don't want the machines in group A to be able to
1113 talk to the machines in group B. You can do this with two
1114 switches, by plugging the machines in group A into one switch and
1115 the machines in group B into the other switch.
1117 If you only have one switch, then you can use VLANs to do the same
1118 thing, by configuring the ports for machines in group A as VLAN
1119 "access ports" for one VLAN and the ports for group B as "access
1120 ports" for a different VLAN. The switch will only forward packets
1121 between ports that are assigned to the same VLAN, so this
1122 effectively subdivides your single switch into two independent
1123 switches, one for each group of machines.
1125 So far we haven't said anything about VLAN headers. With access
1126 ports, like we've described so far, no VLAN header is present in
1127 the Ethernet frame. This means that the machines (or switches)
1128 connected to access ports need not be aware that VLANs are
1129 involved, just like in the case where we use two different physical
1132 Now suppose that you have a whole bunch of switches in your
1133 network, instead of just one, and that some machines in group A are
1134 connected directly to both switches 1 and 2. To allow these
1135 machines to talk to each other, you could add an access port for
1136 group A's VLAN to switch 1 and another to switch 2, and then
1137 connect an Ethernet cable between those ports. That works fine,
1138 but it doesn't scale well as the number of switches and the number
1139 of VLANs increases, because you use up a lot of valuable switch
1140 ports just connecting together your VLANs.
1142 This is where VLAN headers come in. Instead of using one cable and
1143 two ports per VLAN to connect a pair of switches, we configure a
1144 port on each switch as a VLAN "trunk port". Packets sent and
1145 received on a trunk port carry a VLAN header that says what VLAN
1146 the packet belongs to, so that only two ports total are required to
1147 connect the switches, regardless of the number of VLANs in use.
1148 Normally, only switches (either physical or virtual) are connected
1149 to a trunk port, not individual hosts, because individual hosts
1150 don't expect to see a VLAN header in the traffic that they receive.
1152 None of the above discussion says anything about particular VLAN
1153 numbers. This is because VLAN numbers are completely arbitrary.
1154 One must only ensure that a given VLAN is numbered consistently
1155 throughout a network and that different VLANs are given different
1156 numbers. (That said, VLAN 0 is usually synonymous with a packet
1157 that has no VLAN header, and VLAN 4095 is reserved.)
1159 ### Q: VLANs don't work.
1161 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1162 bugs. If you are having problems with VLANs that you suspect to be
1163 driver related, then you have several options:
1165 - Upgrade to Linux 3.3 or later.
1167 - Build and install a fixed version of the particular driver
1168 that is causing trouble, if one is available.
1170 - Use a NIC whose driver does not have VLAN problems.
1172 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
1173 that works around bugs in kernel drivers. To enable VLAN
1174 splinters on interface eth0, use the command:
1176 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1178 For VLAN splinters to be effective, Open vSwitch must know
1179 which VLANs are in use. See the "VLAN splinters" section in
1180 the Interface table in ovs-vswitchd.conf.db(5) for details on
1181 how Open vSwitch infers in-use VLANs.
1183 VLAN splinters increase memory use and reduce performance, so
1184 use them only if needed.
1186 - Apply the "vlan workaround" patch from the XenServer kernel
1187 patch queue, build Open vSwitch against this patched kernel,
1188 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1189 workaround for each interface whose driver is buggy.
1191 (This is a nontrivial exercise, so this option is included
1192 only for completeness.)
1194 It is not always easy to tell whether a Linux kernel driver has
1195 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1196 can help you test. See their manpages for details. Of the two
1197 utilities, ovs-test(8) is newer and more thorough, but
1198 ovs-vlan-test(8) may be easier to use.
1200 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1202 A: Do you have VLANs enabled on the physical switch that OVS is
1203 attached to? Make sure that the port is configured to trunk the
1204 VLAN or VLANs that you are using with OVS.
1206 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1207 and to its destination host, but OVS seems to drop incoming return
1210 A: It's possible that you have the VLAN configured on your physical
1211 switch as the "native" VLAN. In this mode, the switch treats
1212 incoming packets either tagged with the native VLAN or untagged as
1213 part of the native VLAN. It may also send outgoing packets in the
1214 native VLAN without a VLAN tag.
1216 If this is the case, you have two choices:
1218 - Change the physical switch port configuration to tag packets
1219 it forwards to OVS with the native VLAN instead of forwarding
1222 - Change the OVS configuration for the physical port to a
1223 native VLAN mode. For example, the following sets up a
1224 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1226 ovs-vsctl add-br br0
1227 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1229 In this situation, "native-untagged" mode will probably work
1230 equally well. Refer to the documentation for the Port table
1231 in ovs-vswitchd.conf.db(5) for more information.
1233 ### Q: I added a pair of VMs on different VLANs, like this:
1235 ovs-vsctl add-br br0
1236 ovs-vsctl add-port br0 eth0
1237 ovs-vsctl add-port br0 tap0 tag=9
1238 ovs-vsctl add-port br0 tap1 tag=10
1240 but the VMs can't access each other, the external network, or the
1243 A: It is to be expected that the VMs can't access each other. VLANs
1244 are a means to partition a network. When you configured tap0 and
1245 tap1 as access ports for different VLANs, you indicated that they
1246 should be isolated from each other.
1248 As for the external network and the Internet, it seems likely that
1249 the machines you are trying to access are not on VLAN 9 (or 10) and
1250 that the Internet is not available on VLAN 9 (or 10).
1252 ### Q: I added a pair of VMs on the same VLAN, like this:
1254 ovs-vsctl add-br br0
1255 ovs-vsctl add-port br0 eth0
1256 ovs-vsctl add-port br0 tap0 tag=9
1257 ovs-vsctl add-port br0 tap1 tag=9
1259 The VMs can access each other, but not the external network or the
1262 A: It seems likely that the machines you are trying to access in the
1263 external network are not on VLAN 9 and that the Internet is not
1264 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1265 trunk VLAN on the upstream switch port to which eth0 is connected.
1267 ### Q: Can I configure an IP address on a VLAN?
1269 A: Yes. Use an "internal port" configured as an access port. For
1270 example, the following configures IP address 192.168.0.7 on VLAN 9.
1271 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1272 they have an 802.1Q header with VLAN 9. Conversely, traffic
1273 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1276 ovs-vsctl add-br br0
1277 ovs-vsctl add-port br0 eth0
1278 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1279 ifconfig vlan9 192.168.0.7
1281 See also the following question.
1283 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1286 ovs-vsctl add-br br0
1287 ovs-vsctl add-port br0 eth0
1288 ifconfig br0 192.168.0.5
1289 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1290 ifconfig vlan9 192.168.0.9
1292 but other hosts that are only on VLAN 0 can reach the IP address
1293 configured on VLAN 9. What's going on?
1295 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1296 approaches to IP address handling in Internet hosts:
1298 - In the "Strong ES Model", where an ES is a host ("End
1299 System"), an IP address is primarily associated with a
1300 particular interface. The host discards packets that arrive
1301 on interface A if they are destined for an IP address that is
1302 configured on interface B. The host never sends packets from
1303 interface A using a source address configured on interface B.
1305 - In the "Weak ES Model", an IP address is primarily associated
1306 with a host. The host accepts packets that arrive on any
1307 interface if they are destined for any of the host's IP
1308 addresses, even if the address is configured on some
1309 interface other than the one on which it arrived. The host
1310 does not restrict itself to sending packets from an IP
1311 address associated with the originating interface.
1313 Linux uses the weak ES model. That means that when packets
1314 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1315 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1316 address, even though they were not received on vlan9, the network
1319 To simulate the strong ES model on Linux, one may add iptables rule
1320 to filter packets based on source and destination address and
1321 adjust ARP configuration with sysctls.
1323 BSD uses the strong ES model.
1325 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1327 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1328 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1329 implementation of the OpenFlow "normal switching" action. By
1330 default, when Open vSwitch isn't connected to a controller and
1331 nothing has been manually configured in the flow table, all traffic
1332 goes through the "normal switching" action. But, if you set up
1333 OpenFlow flows on your own, through a controller or using ovs-ofctl
1334 or through other means, then you have to implement VLAN handling
1337 You can use "normal switching" as a component of your OpenFlow
1338 actions, e.g. by putting "normal" into the lists of actions on
1339 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1340 controller. In situations where this is not suitable, you can
1341 implement VLAN handling yourself, e.g.:
1343 - If a packet comes in on an access port, and the flow table
1344 needs to send it out on a trunk port, then the flow can add
1345 the appropriate VLAN tag with the "mod_vlan_vid" action.
1347 - If a packet comes in on a trunk port, and the flow table
1348 needs to send it out on an access port, then the flow can
1349 strip the VLAN tag with the "strip_vlan" action.
1351 ### Q: I configured ports on a bridge as access ports with different VLAN
1354 ovs-vsctl add-br br0
1355 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1356 ovs-vsctl add-port br0 eth0
1357 ovs-vsctl add-port br0 tap0 tag=9
1358 ovs-vsctl add-port br0 tap1 tag=10
1360 but the VMs running behind tap0 and tap1 can still communicate,
1361 that is, they are not isolated from each other even though they are
1364 A: Do you have a controller configured on br0 (as the commands above
1365 do)? If so, then this is a variant on the previous question, "My
1366 OpenFlow controller doesn't see the VLANs that I expect," and you
1367 can refer to the answer there for more information.
1369 ### Q: How MAC learning works with VLANs?
1371 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1372 OFPP_NORMAL action. I.e. it logically has separate learning tables
1379 ### Q: What's a VXLAN?
1381 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1382 to solve the scaling challenges of VLAN networks in a multi-tenant
1383 environment. VXLAN is an overlay network which transports an L2 network
1384 over an existing L3 network. For more information on VXLAN, please see
1387 http://tools.ietf.org/html/rfc7348
1389 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1391 A: Open vSwitch currently supports the framing format for packets on the
1392 wire. There is currently no support for the multicast aspects of VXLAN.
1393 To get around the lack of multicast support, it is possible to
1394 pre-provision MAC to IP address mappings either manually or from a
1397 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1400 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1401 is 4789. However, it is possible to configure the destination UDP port
1402 manually on a per-VXLAN tunnel basis. An example of this configuration is
1405 ovs-vsctl add-br br0
1406 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1407 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1408 options:dst_port=8472
1411 Using OpenFlow (Manually or Via Controller)
1412 -------------------------------------------
1414 ### Q: What versions of OpenFlow does Open vSwitch support?
1416 A: The following table lists the versions of OpenFlow supported by
1417 each version of Open vSwitch:
1419 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5
1420 ###============ ===== ===== ===== ===== ===== =====
1421 1.9 and earlier yes --- --- --- --- ---
1422 1.10 yes --- [*] [*] --- ---
1423 1.11 yes --- [*] [*] --- ---
1424 2.0 yes [*] [*] [*] --- ---
1425 2.1 yes [*] [*] [*] --- ---
1426 2.2 yes [*] [*] [*] [%] [*]
1427 2.3 yes yes yes yes [*] [*]
1429 [*] Supported, with one or more missing features.
1430 [%] Experimental, unsafe implementation.
1432 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1433 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1434 1.2, and 1.3 must be enabled manually in ovs-vswitchd. OpenFlow
1435 1.4 and 1.5 are also supported, with missing features, in Open
1436 vSwitch 2.3 and later, but not enabled by default. In any case,
1437 the user may override the default:
1439 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1441 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1443 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1445 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1447 - To enable only OpenFlow 1.0 on bridge br0:
1449 ovs-vsctl set bridge br0 protocols=OpenFlow10
1451 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1452 default. Use the -O option to enable support for later versions of
1453 OpenFlow in ovs-ofctl. For example:
1455 ovs-ofctl -O OpenFlow13 dump-flows br0
1457 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1458 1.4 that could cause crashes. We don't recommend enabling it.)
1460 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1461 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1462 1.5 is solidly implemented, Open vSwitch will enable those version
1465 ### Q: Does Open vSwitch support MPLS?
1467 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1468 these versions can match on MPLS Ethernet types, but they cannot
1469 match, push, or pop MPLS labels, nor can they look past MPLS labels
1470 into the encapsulated packet.
1472 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1473 for MPLS. With the userspace datapath only, these versions can
1474 match, push, or pop a single MPLS label, but they still cannot look
1475 past MPLS labels (even after popping them) into the encapsulated
1476 packet. Kernel datapath support is unchanged from earlier
1479 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1480 label and look past the MPLS label into the encapsulated packet.
1481 Both userspace and kernel datapaths will be supported, but MPLS
1482 processing always happens in userspace either way, so kernel
1483 datapath performance will be disappointing.
1485 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1486 labels and look past the MPLS label into the encapsulated packet.
1487 It will have kernel support for MPLS, yielding improved
1490 ### Q: I'm getting "error type 45250 code 0". What's that?
1492 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1493 vSwitch uses this extension when it must report an error to an
1494 OpenFlow controller but no standard OpenFlow error code is
1497 Open vSwitch logs the errors that it sends to controllers, so the
1498 easiest thing to do is probably to look at the ovs-vswitchd log to
1499 find out what the error was.
1501 If you want to dissect the extended error message yourself, the
1502 format is documented in include/openflow/nicira-ext.h in the Open
1503 vSwitch source distribution. The extended error codes are
1504 documented in lib/ofp-errors.h.
1506 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1507 doesn't actually appear through the OpenFlow connection, even
1508 though I know that it's going through.
1509 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1510 to apply to certain traffic, especially traffic between OVS and
1511 the controller itself.
1513 A: By default, Open vSwitch assumes that OpenFlow controllers are
1514 connected "in-band", that is, that the controllers are actually
1515 part of the network that is being controlled. In in-band mode,
1516 Open vSwitch sets up special "hidden" flows to make sure that
1517 traffic can make it back and forth between OVS and the controllers.
1518 These hidden flows are higher priority than any flows that can be
1519 set up through OpenFlow, and they are not visible through normal
1520 OpenFlow flow table dumps.
1522 Usually, the hidden flows are desirable and helpful, but
1523 occasionally they can cause unexpected behavior. You can view the
1524 full OpenFlow flow table, including hidden flows, on bridge br0
1527 ovs-appctl bridge/dump-flows br0
1529 to help you debug. The hidden flows are those with priorities
1530 greater than 65535 (the maximum priority that can be set with
1533 The DESIGN file at the top level of the Open vSwitch source
1534 distribution describes the in-band model in detail.
1536 If your controllers are not actually in-band (e.g. they are on
1537 localhost via 127.0.0.1, or on a separate network), then you should
1538 configure your controllers in "out-of-band" mode. If you have one
1539 controller on bridge br0, then you can configure out-of-band mode
1542 ovs-vsctl set controller br0 connection-mode=out-of-band
1544 ### Q: I configured all my controllers for out-of-band control mode but
1545 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1547 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1548 set-manager"). By default, Open vSwitch assumes that managers need
1549 in-band rules set up on every bridge. You can disable these rules
1552 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1554 This actually disables in-band control entirely for the bridge, as
1555 if all the bridge's controllers were configured for out-of-band
1558 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1560 A: See answer under "VLANs", above.
1562 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1563 but I got a funny message like this:
1565 ofp_util|INFO|normalization changed ofp_match, details:
1566 ofp_util|INFO| pre: nw_dst=192.168.0.1
1569 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1570 match had disappeared, so that the flow ends up matching every
1573 A: The term "normalization" in the log message means that a flow
1574 cannot match on an L3 field without saying what L3 protocol is in
1575 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1576 so the L3 field match was dropped.
1578 In this case, the L3 protocol could be IP or ARP. A correct
1579 command for each possibility is, respectively:
1581 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1585 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1587 Similarly, a flow cannot match on an L4 field without saying what
1588 L4 protocol is in use. For example, the flow match "tp_src=1234"
1589 is, by itself, meaningless and will be ignored. Instead, to match
1590 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1591 source port 1234, write "udp,tp_src=1234".
1593 ### Q: How can I figure out the OpenFlow port number for a given port?
1595 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1596 respond with an OFPT_FEATURES_REPLY that, among other information,
1597 includes a mapping between OpenFlow port names and numbers. From a
1598 command prompt, "ovs-ofctl show br0" makes such a request and
1599 prints the response for switch br0.
1601 The Interface table in the Open vSwitch database also maps OpenFlow
1602 port names to numbers. To print the OpenFlow port number
1603 associated with interface eth0, run:
1605 ovs-vsctl get Interface eth0 ofport
1607 You can print the entire mapping with:
1609 ovs-vsctl -- --columns=name,ofport list Interface
1611 but the output mixes together interfaces from all bridges in the
1612 database, so it may be confusing if more than one bridge exists.
1614 In the Open vSwitch database, ofport value -1 means that the
1615 interface could not be created due to an error. (The Open vSwitch
1616 log should indicate the reason.) ofport value [] (the empty set)
1617 means that the interface hasn't been created yet. The latter is
1618 normally an intermittent condition (unless ovs-vswitchd is not
1621 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1622 run "ovs-dpctl dump-flows" I don't see them.
1624 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1625 won't display the information that you want. You want to use
1626 "ovs-ofctl dump-flows" instead.
1628 ### Q: It looks like each of the interfaces in my bonded port shows up
1629 as an individual OpenFlow port. Is that right?
1631 A: Yes, Open vSwitch makes individual bond interfaces visible as
1632 OpenFlow ports, rather than the bond as a whole. The interfaces
1633 are treated together as a bond for only a few purposes:
1635 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1636 controller is not configured, this happens implicitly to
1639 - Mirrors configured for output to a bonded port.
1641 It would make a lot of sense for Open vSwitch to present a bond as
1642 a single OpenFlow port. If you want to contribute an
1643 implementation of such a feature, please bring it up on the Open
1644 vSwitch development mailing list at dev@openvswitch.org.
1646 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1647 multiple hosts, and other components. The behavior isn't what I
1650 A: To debug network behavior problems, trace the path of a packet,
1651 hop-by-hop, from its origin in one host to a remote host. If
1652 that's correct, then trace the path of the response packet back to
1655 The open source tool called "plotnetcfg" can help to understand the
1656 relationship between the networking devices on a single host.
1658 Usually a simple ICMP echo request and reply ("ping") packet is
1659 good enough. Start by initiating an ongoing "ping" from the origin
1660 host to a remote host. If you are tracking down a connectivity
1661 problem, the "ping" will not display any successful output, but
1662 packets are still being sent. (In this case the packets being sent
1663 are likely ARP rather than ICMP.)
1665 Tools available for tracing include the following:
1667 - "tcpdump" and "wireshark" for observing hops across network
1668 devices, such as Open vSwitch internal devices and physical
1671 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1672 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1673 These tools allow one to observe the actions being taken on
1674 packets in ongoing flows.
1676 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1677 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1678 documentation, and "Why are there so many different ways to
1679 dump flows?" above for some background.
1681 - "ovs-appctl ofproto/trace" to observe the logic behind how
1682 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1683 documentation. You can out more details about a given flow
1684 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1685 a flow from the output into an "ovs-appctl ofproto/trace"
1688 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1689 observe what goes on at these physical hops.
1691 Starting at the origin of a given packet, observe the packet at
1692 each hop in turn. For example, in one plausible scenario, you
1695 1. "tcpdump" the "eth" interface through which an ARP egresses
1696 a VM, from inside the VM.
1698 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1699 ingresses the host machine.
1701 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1702 the host interface through which the ARP egresses the
1703 physical machine. You may need to use "ovs-dpctl show" to
1704 interpret the port numbers. If the output seems surprising,
1705 you can use "ovs-appctl ofproto/trace" to observe details of
1706 how ovs-vswitchd determined the actions in the "ovs-dpctl
1709 4. "tcpdump" the "eth" interface through which the ARP egresses
1710 the physical machine.
1712 5. "tcpdump" the "eth" interface through which the ARP
1713 ingresses the physical machine, at the remote host that
1716 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1717 remote host that receives the ARP and observe the VM "vif"
1718 or "tap" interface to which the flow is directed. Again,
1719 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1721 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1724 8. "tcpdump" the "eth" interface through which the ARP
1725 ingresses a VM, from inside the VM.
1727 It is likely that during one of these steps you will figure out the
1728 problem. If not, then follow the ARP reply back to the origin, in
1731 ### Q: How do I make a flow drop packets?
1733 A: To drop a packet is to receive it without forwarding it. OpenFlow
1734 explicitly specifies forwarding actions. Thus, a flow with an
1735 empty set of actions does not forward packets anywhere, causing
1736 them to be dropped. You can specify an empty set of actions with
1737 "actions=" on the ovs-ofctl command line. For example:
1739 ovs-ofctl add-flow br0 priority=65535,actions=
1741 would cause every packet entering switch br0 to be dropped.
1743 You can write "drop" explicitly if you like. The effect is the
1744 same. Thus, the following command also causes every packet
1745 entering switch br0 to be dropped:
1747 ovs-ofctl add-flow br0 priority=65535,actions=drop
1749 "drop" is not an action, either in OpenFlow or Open vSwitch.
1750 Rather, it is only a way to say that there are no actions.
1752 ### Q: I added a flow to send packets out the ingress port, like this:
1754 ovs-ofctl add-flow br0 in_port=2,actions=2
1756 but OVS drops the packets instead.
1758 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1759 out its ingress port. The rationale is that dropping these packets
1760 makes it harder to loop the network. Sometimes this behavior can
1761 even be convenient, e.g. it is often the desired behavior in a flow
1762 that forwards a packet to several ports ("floods" the packet).
1764 Sometimes one really needs to send a packet out its ingress port
1765 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1766 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1768 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1770 This also works in some circumstances where the flow doesn't match
1771 on the input port. For example, if you know that your switch has
1772 five ports numbered 2 through 6, then the following will send every
1773 received packet out every port, even its ingress port:
1775 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1779 ovs-ofctl add-flow br0 actions=all,in_port
1781 Sometimes, in complicated flow tables with multiple levels of
1782 "resubmit" actions, a flow needs to output to a particular port
1783 that may or may not be the ingress port. It's difficult to take
1784 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1785 provides, as an OpenFlow extension, the ability to modify the
1786 in_port field. Whatever value is currently in the in_port field is
1787 the port to which outputs will be dropped, as well as the
1788 destination for OFPP_IN_PORT. This means that the following will
1789 reliably output to port 2 or to ports 2 through 6, respectively:
1791 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1792 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1794 If the input port is important, then one may save and restore it on
1797 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1798 load:0->NXM_OF_IN_PORT[],\
1800 pop:NXM_OF_IN_PORT[]
1802 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1803 on port 2. I set up flows to forward only traffic destined to the
1804 other host and drop other traffic, like this:
1806 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1807 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1808 priority=0,actions=drop
1810 But it doesn't work--I don't get any connectivity when I do this.
1813 A: These flows drop the ARP packets that IP hosts use to establish IP
1814 connectivity over Ethernet. To solve the problem, add flows to
1815 allow ARP to pass between the hosts:
1817 priority=5,in_port=1,arp,actions=2
1818 priority=5,in_port=2,arp,actions=1
1820 This issue can manifest other ways, too. The following flows that
1821 match on Ethernet addresses instead of IP addresses will also drop
1822 ARP packets, because ARP requests are broadcast instead of being
1823 directed to a specific host:
1825 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1826 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1827 priority=0,actions=drop
1829 The solution already described above will also work in this case.
1830 It may be better to add flows to allow all multicast and broadcast
1833 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1834 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1836 ### Q: My bridge disconnects from my controller on add-port/del-port.
1838 A: Reconfiguring your bridge can change your bridge's datapath-id because
1839 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1840 In that case, Open vSwitch disconnects from controllers because there's
1841 no graceful way to notify controllers about the change of datapath-id.
1843 To avoid the behaviour, you can configure datapath-id manually.
1845 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1847 ### Q: My controller is getting errors about "buffers". What's going on?
1849 A: When a switch sends a packet to an OpenFlow controller using a
1850 "packet-in" message, it can also keep a copy of that packet in a
1851 "buffer", identified by a 32-bit integer "buffer_id". There are
1852 two advantages to buffering. First, when the controller wants to
1853 tell the switch to do something with the buffered packet (with a
1854 "packet-out" OpenFlow request), it does not need to send another
1855 copy of the packet back across the OpenFlow connection, which
1856 reduces the bandwidth cost of the connection and improves latency.
1857 This enables the second advantage: the switch can optionally send
1858 only the first part of the packet to the controller (assuming that
1859 the switch only needs to look at the first few bytes of the
1860 packet), further reducing bandwidth and improving latency.
1862 However, buffering introduces some issues of its own. First, any
1863 switch has limited resources, so if the controller does not use a
1864 buffered packet, the switch has to decide how long to keep it
1865 buffered. When many packets are sent to a controller and buffered,
1866 Open vSwitch can discard buffered packets that the controller has
1867 not used after as little as 5 seconds. This means that
1868 controllers, if they make use of packet buffering, should use the
1869 buffered packets promptly. (This includes sending a "packet-out"
1870 with no actions if the controller does not want to do anything with
1871 a buffered packet, to clear the packet buffer and effectively
1874 Second, packet buffers are one-time-use, meaning that a controller
1875 cannot use a single packet buffer in two or more "packet-out"
1876 commands. Open vSwitch will respond with an error to the second
1877 and subsequent "packet-out"s in such a case.
1879 Finally, a common error early in controller development is to try
1880 to use buffer_id 0 in a "packet-out" message as if 0 represented
1881 "no buffered packet". This is incorrect usage: the buffer_id with
1882 this meaning is actually 0xffffffff.
1884 ovs-vswitchd(8) describes some details of Open vSwitch packet
1885 buffering that the OpenFlow specification requires implementations
1892 ### Q: How do I implement a new OpenFlow message?
1894 A: Add your new message to "enum ofpraw" and "enum ofptype" in
1895 lib/ofp-msgs.h, following the existing pattern. Then recompile and
1896 fix all of the new warnings, implementing new functionality for the
1897 new message as needed. (If you configure with --enable-Werror, as
1898 described in [INSTALL.md], then it is impossible to miss any warnings.)
1900 If you need to add an OpenFlow vendor extension message for a
1901 vendor that doesn't yet have any extension messages, then you will
1902 also need to edit build-aux/extract-ofp-msgs.
1904 ### Q: How do I add support for a new field or header?
1906 A: Add new members for your field to "struct flow" in lib/flow.h, and
1907 add new enumerations for your new field to "enum mf_field_id" in
1908 lib/meta-flow.h, following the existing pattern. Also, add support
1909 to miniflow_extract() in lib/flow.c for extracting your new field
1910 from a packet into struct miniflow. Then recompile and fix all of
1911 the new warnings, implementing new functionality for the new field
1912 or header as needed. (If you configure with --enable-Werror, as
1913 described in [INSTALL.md], then it is impossible to miss any
1916 If you want kernel datapath support for your new field, you also
1917 need to modify the kernel module for the operating systems you are
1918 interested in. This isn't mandatory, since fields understood only
1919 by userspace work too (with a performance penalty), so it's
1920 reasonable to start development without it. If you implement
1921 kernel module support for Linux, then the Linux kernel "netdev"
1922 mailing list is the place to submit that support first; please read
1923 up on the Linux kernel development process separately. The Windows
1924 datapath kernel module support, on the other hand, is maintained
1925 within the OVS tree, so patches for that can go directly to
1928 ### Q: How do I add support for a new OpenFlow action?
1930 A: Add your new action to "enum ofp_raw_action_type" in
1931 lib/ofp-actions.c, following the existing pattern. Then recompile
1932 and fix all of the new warnings, implementing new functionality for
1933 the new action as needed. (If you configure with --enable-Werror,
1934 as described in [INSTALL.md], then it is impossible to miss any
1937 If you need to add an OpenFlow vendor extension action for a vendor
1938 that doesn't yet have any extension actions, then you will also
1939 need to edit build-aux/extract-ofp-actions.
1945 bugs@openvswitch.org
1946 http://openvswitch.org/
1948 [PORTING.md]:PORTING.md
1949 [WHY-OVS.md]:WHY-OVS.md
1950 [INSTALL.md]:INSTALL.md
1951 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
1952 [INSTALL.DPDK.md]:INSTALL.DPDK.md