1 Frequently Asked Questions
2 ==========================
4 Open vSwitch <http://openvswitch.org>
9 ### Q: What is Open vSwitch?
11 A: Open vSwitch is a production quality open source software switch
12 designed to be used as a vswitch in virtualized server
13 environments. A vswitch forwards traffic between different VMs on
14 the same physical host and also forwards traffic between VMs and
15 the physical network. Open vSwitch supports standard management
16 interfaces (e.g. sFlow, NetFlow, IPFIX, RSPAN, CLI), and is open to
17 programmatic extension and control using OpenFlow and the OVSDB
20 Open vSwitch as designed to be compatible with modern switching
21 chipsets. This means that it can be ported to existing high-fanout
22 switches allowing the same flexible control of the physical
23 infrastructure as the virtual infrastructure. It also means that
24 Open vSwitch will be able to take advantage of on-NIC switching
25 chipsets as their functionality matures.
27 ### Q: What virtualization platforms can use Open vSwitch?
29 A: Open vSwitch can currently run on any Linux-based virtualization
30 platform (kernel 3.10 and newer), including: KVM, VirtualBox, Xen,
31 Xen Cloud Platform, XenServer. As of Linux 3.3 it is part of the
32 mainline kernel. The bulk of the code is written in platform-
33 independent C and is easily ported to other environments. We welcome
34 inquires about integrating Open vSwitch with other virtualization
37 ### Q: How can I try Open vSwitch?
39 A: The Open vSwitch source code can be built on a Linux system. You can
40 build and experiment with Open vSwitch on any Linux machine.
41 Packages for various Linux distributions are available on many
42 platforms, including: Debian, Ubuntu, Fedora.
44 You may also download and run a virtualization platform that already
45 has Open vSwitch integrated. For example, download a recent ISO for
46 XenServer or Xen Cloud Platform. Be aware that the version
47 integrated with a particular platform may not be the most recent Open
50 ### Q: Does Open vSwitch only work on Linux?
52 A: No, Open vSwitch has been ported to a number of different operating
53 systems and hardware platforms. Most of the development work occurs
54 on Linux, but the code should be portable to any POSIX system. We've
55 seen Open vSwitch ported to a number of different platforms,
56 including FreeBSD, Windows, and even non-POSIX embedded systems.
58 By definition, the Open vSwitch Linux kernel module only works on
59 Linux and will provide the highest performance. However, a userspace
60 datapath is available that should be very portable.
62 ### Q: What's involved with porting Open vSwitch to a new platform or switching ASIC?
64 A: The [PORTING.md] document describes how one would go about
65 porting Open vSwitch to a new operating system or hardware platform.
67 ### Q: Why would I use Open vSwitch instead of the Linux bridge?
69 A: Open vSwitch is specially designed to make it easier to manage VM
70 network configuration and monitor state spread across many physical
71 hosts in dynamic virtualized environments. Please see
72 [WHY-OVS.md] for a more detailed description of how Open vSwitch
73 relates to the Linux Bridge.
75 ### Q: How is Open vSwitch related to distributed virtual switches like the VMware vNetwork distributed switch or the Cisco Nexus 1000V?
77 A: Distributed vswitch applications (e.g., VMware vNetwork distributed
78 switch, Cisco Nexus 1000V) provide a centralized way to configure and
79 monitor the network state of VMs that are spread across many physical
80 hosts. Open vSwitch is not a distributed vswitch itself, rather it
81 runs on each physical host and supports remote management in a way
82 that makes it easier for developers of virtualization/cloud
83 management platforms to offer distributed vswitch capabilities.
85 To aid in distribution, Open vSwitch provides two open protocols that
86 are specially designed for remote management in virtualized network
87 environments: OpenFlow, which exposes flow-based forwarding state,
88 and the OVSDB management protocol, which exposes switch port state.
89 In addition to the switch implementation itself, Open vSwitch
90 includes tools (ovs-ofctl, ovs-vsctl) that developers can script and
91 extend to provide distributed vswitch capabilities that are closely
92 integrated with their virtualization management platform.
94 ### Q: Why doesn't Open vSwitch support distribution?
96 A: Open vSwitch is intended to be a useful component for building
97 flexible network infrastructure. There are many different approaches
98 to distribution which balance trade-offs between simplicity,
99 scalability, hardware compatibility, convergence times, logical
100 forwarding model, etc. The goal of Open vSwitch is to be able to
101 support all as a primitive building block rather than choose a
102 particular point in the distributed design space.
104 ### Q: How can I contribute to the Open vSwitch Community?
106 A: You can start by joining the mailing lists and helping to answer
107 questions. You can also suggest improvements to documentation. If
108 you have a feature or bug you would like to work on, send a mail to
109 one of the mailing lists:
111 http://openvswitch.org/mlists/
113 ### Q: Why can I no longer connect to my OpenFlow controller or OVSDB manager?
115 A: Starting in OVS 2.4, we switched the default ports to the
116 IANA-specified port numbers for OpenFlow (6633->6653) and OVSDB
117 (6632->6640). We recommend using these port numbers, but if you
118 cannot, all the programs allow overriding the default port. See the
119 appropriate man page.
125 ### Q: What does it mean for an Open vSwitch release to be LTS (long-term support)?
127 A: All official releases have been through a comprehensive testing
128 process and are suitable for production use. Planned releases will
129 occur several times a year. If a significant bug is identified in an
130 LTS release, we will provide an updated release that includes the
131 fix. Releases that are not LTS may not be fixed and may just be
132 supplanted by the next major release. The current LTS release is
135 ### Q: What Linux kernel versions does each Open vSwitch release work with?
137 A: The following table lists the Linux kernel versions against which the
138 given versions of the Open vSwitch kernel module will successfully
139 build. The Linux kernel versions are upstream kernel versions, so
140 Linux kernels modified from the upstream sources may not build in
141 some cases even if they are based on a supported version. This is
142 most notably true of Red Hat Enterprise Linux (RHEL) kernels, which
143 are extensively modified from upstream.
145 | Open vSwitch | Linux kernel
146 |:------------:|:-------------:
147 | 1.4.x | 2.6.18 to 3.2
148 | 1.5.x | 2.6.18 to 3.2
149 | 1.6.x | 2.6.18 to 3.2
150 | 1.7.x | 2.6.18 to 3.3
151 | 1.8.x | 2.6.18 to 3.4
152 | 1.9.x | 2.6.18 to 3.8
153 | 1.10.x | 2.6.18 to 3.8
154 | 1.11.x | 2.6.18 to 3.8
155 | 2.0.x | 2.6.32 to 3.10
156 | 2.1.x | 2.6.32 to 3.11
157 | 2.3.x | 2.6.32 to 3.14
158 | 2.4.x | 2.6.32 to 4.0
159 | 2.5.x | 2.6.32 to 4.3
160 | 2.6.x | 3.10 to 4.3
162 Open vSwitch userspace should also work with the Linux kernel module
163 built into Linux 3.3 and later.
165 Open vSwitch userspace is not sensitive to the Linux kernel version.
166 It should build against almost any kernel, certainly against 2.6.32
169 ### Q: Are all features available with all datapaths?
171 A: Open vSwitch supports different datapaths on different platforms. Each
172 datapath has a different feature set: the following tables try to summarize
177 * *Linux upstream*: The datapath implemented by the kernel module shipped
178 with Linux upstream. Since features have been gradually
179 introduced into the kernel, the table mentions the first
180 Linux release whose OVS module supports the feature.
182 * *Linux OVS tree*: The datapath implemented by the Linux kernel module
183 distributed with the OVS source tree.
185 * *Userspace*: Also known as DPDK, dpif-netdev or dummy datapath. It is the
186 only datapath that works on NetBSD, FreeBSD and Mac OSX.
188 * *Hyper-V*: Also known as the Windows datapath.
190 The following table lists the datapath supported features from
191 an Open vSwitch user's perspective.
193 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
194 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
195 NAT | 4.6 | NO | NO | NO |
196 Connection tracking | 4.3 | YES | NO | PARTIAL |
197 Tunnel - LISP | NO | YES | NO | NO |
198 Tunnel - STT | NO | YES | NO | YES |
199 Tunnel - GRE | 3.11 | YES | YES | YES |
200 Tunnel - VXLAN | 3.12 | YES | YES | YES |
201 Tunnel - Geneve | 3.18 | YES | YES | NO |
202 QoS - Policing | YES | YES | NO | NO |
203 QoS - Shaping | YES | YES | NO | NO |
204 sFlow | YES | YES | YES | NO |
205 IPFIX | 3.10 | YES | YES | NO |
206 Set action | YES | YES | YES | PARTIAL |
207 NIC Bonding | YES | YES | YES | NO |
208 Multiple VTEPs | YES | YES | YES | NO |
211 * Only a limited set of flow fields is modifiable via the set action by the
213 * The Hyper-V datapath only supports one physical NIC per datapath. This is
214 why bonding is not supported.
215 * The Hyper-V datapath can have at most one IP address configured as a
218 The following table lists features that do not *directly* impact an
219 Open vSwitch user, e.g. because their absence can be hidden by the ofproto
220 layer (usually this comes with a performance penalty).
222 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
223 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
224 SCTP flows | 3.12 | YES | YES | YES |
225 MPLS | 3.19 | YES | YES | YES |
226 UFID | 4.0 | YES | YES | NO |
227 Megaflows | 3.12 | YES | YES | NO |
228 Masked set action | 4.0 | YES | YES | NO |
229 Recirculation | 3.19 | YES | YES | YES |
230 TCP flags matching | 3.13 | YES | YES | NO |
231 Validate flow actions | YES | YES | N/A | NO |
232 Multiple datapaths | YES | YES | YES | NO |
233 Tunnel TSO - STT | N/A | YES | NO | YES |
235 ### Q: What DPDK version does each Open vSwitch release work with?
237 A: The following table lists the DPDK version against which the
238 given versions of Open vSwitch will successfully build.
240 | Open vSwitch | DPDK
241 |:------------:|:-----:
248 ### Q: I get an error like this when I configure Open vSwitch:
250 configure: error: Linux kernel in <dir> is version <x>, but
251 version newer than <y> is not supported (please refer to the
256 A: You have the following options:
258 - Use the Linux kernel module supplied with the kernel that you are
259 using. (See also the following FAQ.)
261 - If there is a newer released version of Open vSwitch, consider
262 building that one, because it may support the kernel that you are
263 building against. (To find out, consult the table in the
266 - The Open vSwitch "master" branch may support the kernel that you
267 are using, so consider building the kernel module from "master".
269 All versions of Open vSwitch userspace are compatible with all
270 versions of the Open vSwitch kernel module, so you do not have to
271 use the kernel module from one source along with the userspace
272 programs from the same source.
274 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
276 A: The kernel module in upstream Linux does not include support for
277 LISP. Work is in progress to add support for LISP to the upstream
278 Linux version of the Open vSwitch kernel module. For now, if you
279 need this feature, use the kernel module from the Open vSwitch
280 distribution instead of the upstream Linux kernel module.
282 Certain features require kernel support to function or to have
283 reasonable performance. If the ovs-vswitchd log file indicates that
284 a feature is not supported, consider upgrading to a newer upstream
285 Linux release or using the kernel module paired with the userspace
288 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
290 A: Support for tunnels was added to the upstream Linux kernel module
291 after the rest of Open vSwitch. As a result, some kernels may contain
292 support for Open vSwitch but not tunnels. The minimum kernel version
293 that supports each tunnel protocol is:
295 | Protocol | Linux Kernel
296 |:--------:|:-------------:
300 | LISP | <not upstream>
301 | STT | <not upstream>
303 If you are using a version of the kernel that is older than the one
304 listed above, it is still possible to use that tunnel protocol. However,
305 you must compile and install the kernel module included with the Open
306 vSwitch distribution rather than the one on your machine. If problems
307 persist after doing this, check to make sure that the module that is
308 loaded is the one you expect.
310 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
312 A: Generating outer UDP checksums requires kernel support that was not
313 part of the initial implementation of these protocols. If using the
314 upstream Linux Open vSwitch module, you must use kernel 4.0 or
315 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
316 support UDP checksums.
318 ### Q: What features are not available when using the userspace datapath?
320 A: Tunnel virtual ports are not supported, as described in the
321 previous answer. It is also not possible to use queue-related
322 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
323 may not be transmitted.
325 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
327 In general, the Open vSwitch userspace should be used with the
328 kernel version included in the same release or with the version
329 from upstream Linux. However, when upgrading between two releases
330 of Open vSwitch it is best to migrate userspace first to reduce
331 the possibility of incompatibilities.
333 ### Q: What happened to the bridge compatibility feature?
335 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
336 When it was enabled, Open vSwitch imitated the interface of the
337 Linux kernel "bridge" module. This allowed users to drop Open
338 vSwitch into environments designed to use the Linux kernel bridge
339 module without adapting the environment to use Open vSwitch.
341 Open vSwitch 1.10 and later do not support bridge compatibility.
342 The feature was dropped because version 1.10 adopted a new internal
343 architecture that made bridge compatibility difficult to maintain.
344 Now that many environments use OVS directly, it would be rarely
347 To use bridge compatibility, install OVS 1.9 or earlier, including
348 the accompanying kernel modules (both the main and bridge
349 compatibility modules), following the instructions that come with
350 the release. Be sure to start the ovs-brcompatd daemon.
356 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
358 A: In networking, the terms "bridge" and "switch" are synonyms. Open
359 vSwitch implements an Ethernet switch, which means that it is also
362 ### Q: What's a VLAN?
364 A: See the "VLAN" section below.
370 ### Q: How do I configure a port as an access port?
372 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
373 the following commands configure br0 with eth0 as a trunk port (the
374 default) and tap0 as an access port for VLAN 9:
377 ovs-vsctl add-port br0 eth0
378 ovs-vsctl add-port br0 tap0 tag=9
380 If you want to configure an already added port as an access port,
381 use "ovs-vsctl set", e.g.:
383 ovs-vsctl set port tap0 tag=9
385 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
387 A: The following commands configure br0 with eth0 and tap0 as trunk
388 ports. All traffic coming in or going out on eth0 or tap0 is also
389 mirrored to tap1; any traffic arriving on tap1 is dropped:
392 ovs-vsctl add-port br0 eth0
393 ovs-vsctl add-port br0 tap0
394 ovs-vsctl add-port br0 tap1 \
395 -- --id=@p get port tap1 \
396 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
397 -- set bridge br0 mirrors=@m
399 To later disable mirroring, run:
401 ovs-vsctl clear bridge br0 mirrors
403 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
405 A: Yes. How you configure it depends on what you mean by "promiscuous
408 - Conventionally, "promiscuous mode" is a feature of a network
409 interface card. Ordinarily, a NIC passes to the CPU only the
410 packets actually destined to its host machine. It discards
411 the rest to avoid wasting memory and CPU cycles. When
412 promiscuous mode is enabled, however, it passes every packet
413 to the CPU. On an old-style shared-media or hub-based
414 network, this allows the host to spy on all packets on the
415 network. But in the switched networks that are almost
416 everywhere these days, promiscuous mode doesn't have much
417 effect, because few packets not destined to a host are
418 delivered to the host's NIC.
420 This form of promiscuous mode is configured in the guest OS of
421 the VMs on your bridge, e.g. with "ifconfig".
423 - The VMware vSwitch uses a different definition of "promiscuous
424 mode". When you configure promiscuous mode on a VMware vNIC,
425 the vSwitch sends a copy of every packet received by the
426 vSwitch to that vNIC. That has a much bigger effect than just
427 enabling promiscuous mode in a guest OS. Rather than getting
428 a few stray packets for which the switch does not yet know the
429 correct destination, the vNIC gets every packet. The effect
430 is similar to replacing the vSwitch by a virtual hub.
432 This "promiscuous mode" is what switches normally call "port
433 mirroring" or "SPAN". For information on how to configure
434 SPAN, see "How do I configure a port as a SPAN port, that is,
435 enable mirroring of all traffic to that port?"
437 ### Q: How do I configure a DPDK port as an access port?
439 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
441 If your version is DPDK-enabled it will support the other-config:dpdk-init
442 configuration in the database and will display lines with "EAL:..."
443 during startup when other_config:dpdk-init is set to 'true'.
445 Secondly, when adding a DPDK port, unlike a system port, the
446 type for the interface must be specified. For example;
449 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
451 Finally, it is required that DPDK port names begin with 'dpdk'.
453 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
456 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
458 A: The following commands configure br0 with eth0 as a trunk port and
459 tap0 as an access port for VLAN 10. All traffic coming in or going
460 out on tap0, as well as traffic coming in or going out on eth0 in
461 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
462 VLAN 10, in cases where one is present, is dropped as part of
466 ovs-vsctl add-port br0 eth0
467 ovs-vsctl add-port br0 tap0 tag=10
469 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
471 -- set bridge br0 mirrors=@m
473 To later disable mirroring, run:
475 ovs-vsctl clear bridge br0 mirrors
477 Mirroring to a VLAN can disrupt a network that contains unmanaged
478 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
479 GRE tunnel has fewer caveats than mirroring to a VLAN and should
480 generally be preferred.
482 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
484 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
485 of the specified output-vlan. This loss of information may make
486 the mirrored traffic too hard to interpret.
488 To mirror multiple VLANs, use the commands above, but specify a
489 comma-separated list of VLANs as the value for select-vlan. To
490 mirror every VLAN, use the commands above, but omit select-vlan and
493 When a packet arrives on a VLAN that is used as a mirror output
494 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
495 floods the packet across all the ports for which the mirror output
496 VLAN is configured. (If an OpenFlow controller is in use, then it
497 can override this behavior through the flow table.) If OVS is used
498 as an intermediate switch, rather than an edge switch, this ensures
499 that the RSPAN traffic is distributed through the network.
501 Mirroring to a VLAN can disrupt a network that contains unmanaged
502 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
503 GRE tunnel has fewer caveats than mirroring to a VLAN and should
504 generally be preferred.
506 ### Q: How do I configure mirroring of all traffic to a GRE tunnel?
508 A: The following commands configure br0 with eth0 and tap0 as trunk
509 ports. All traffic coming in or going out on eth0 or tap0 is also
510 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
511 traffic arriving on gre0 is dropped:
514 ovs-vsctl add-port br0 eth0
515 ovs-vsctl add-port br0 tap0
516 ovs-vsctl add-port br0 gre0 \
517 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
518 -- --id=@p get port gre0 \
519 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
520 -- set bridge br0 mirrors=@m
522 To later disable mirroring and destroy the GRE tunnel:
524 ovs-vsctl clear bridge br0 mirrors
525 ovs-vsctl del-port br0 gre0
527 ### Q: Does Open vSwitch support ERSPAN?
529 A: No. ERSPAN is an undocumented proprietary protocol. As an
530 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
533 ### Q: How do I connect two bridges?
535 A: First, why do you want to do this? Two connected bridges are not
536 much different from a single bridge, so you might as well just have
537 a single bridge with all your ports on it.
539 If you still want to connect two bridges, you can use a pair of
540 patch ports. The following example creates bridges br0 and br1,
541 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
542 and br1 with a pair of patch ports.
545 ovs-vsctl add-port br0 eth0
546 ovs-vsctl add-port br0 tap0
548 ovs-vsctl add-port br1 tap1
550 -- add-port br0 patch0 \
551 -- set interface patch0 type=patch options:peer=patch1 \
552 -- add-port br1 patch1 \
553 -- set interface patch1 type=patch options:peer=patch0
555 Bridges connected with patch ports are much like a single bridge.
556 For instance, if the example above also added eth1 to br1, and both
557 eth0 and eth1 happened to be connected to the same next-hop switch,
558 then you could loop your network just as you would if you added
559 eth0 and eth1 to the same bridge (see the "Configuration Problems"
560 section below for more information).
562 If you are using Open vSwitch 1.9 or an earlier version, then you
563 need to be using the kernel module bundled with Open vSwitch rather
564 than the one that is integrated into Linux 3.3 and later, because
565 Open vSwitch 1.9 and earlier versions need kernel support for patch
566 ports. This also means that in Open vSwitch 1.9 and earlier, patch
567 ports will not work with the userspace datapath, only with the
570 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
572 A: Open vSwitch does not support such a configuration.
573 Bridges always have their local ports.
576 Implementation Details
577 ----------------------
579 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
581 A: Open vSwitch uses different kinds of flows for different purposes:
583 - OpenFlow flows are the most important kind of flow. OpenFlow
584 controllers use these flows to define a switch's policy.
585 OpenFlow flows support wildcards, priorities, and multiple
588 When in-band control is in use, Open vSwitch sets up a few
589 "hidden" flows, with priority higher than a controller or the
590 user can configure, that are not visible via OpenFlow. (See
591 the "Controller" section of the FAQ for more information
594 - The Open vSwitch software switch implementation uses a second
595 kind of flow internally. These flows, called "datapath" or
596 "kernel" flows, do not support priorities and comprise only a
597 single table, which makes them suitable for caching. (Like
598 OpenFlow flows, datapath flows do support wildcarding, in Open
599 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
600 also support different actions and number ports differently.
602 Datapath flows are an implementation detail that is subject to
603 change in future versions of Open vSwitch. Even with the
604 current version of Open vSwitch, hardware switch
605 implementations do not necessarily use this architecture.
607 Users and controllers directly control only the OpenFlow flow
608 table. Open vSwitch manages the datapath flow table itself, so
609 users should not normally be concerned with it.
611 ### Q: Why are there so many different ways to dump flows?
613 A: Open vSwitch has two kinds of flows (see the previous question), so
614 it has commands with different purposes for dumping each kind of
617 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
618 hidden flows. This is the most commonly useful form of flow
619 dump. (Unlike the other commands, this should work with any
620 OpenFlow switch, not just Open vSwitch.)
622 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
623 including hidden flows. This is occasionally useful for
624 troubleshooting suspected issues with in-band control.
626 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
627 entries for a Linux kernel-based datapath. In Open vSwitch
628 1.10 and later, ovs-vswitchd merges multiple switches into a
629 single datapath, so it will show all the flows on all your
630 kernel-based switches. This command can occasionally be
631 useful for debugging.
633 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
634 dumps datapath flows for only the specified bridge, regardless
637 ### Q: How does multicast snooping works with VLANs?
639 A: Open vSwitch maintains snooping tables for each VLAN.
641 ### Q: Can OVS populate the kernel flow table in advance instead of in reaction to packets?
643 A: No. There are several reasons:
645 - Kernel flows are not as sophisticated as OpenFlow flows, which
646 means that some OpenFlow policies could require a large number of
647 kernel flows. The "conjunctive match" feature is an extreme
648 example: the number of kernel flows it requires is the product of
649 the number of flows in each dimension.
651 - With multiple OpenFlow flow tables and simple sets of actions, the
652 number of kernel flows required can be as large as the product of
653 the number of flows in each dimension. With more sophisticated
654 actions, the number of kernel flows could be even larger.
656 - Open vSwitch is designed so that any version of OVS userspace
657 interoperates with any version of the OVS kernel module. This
658 forward and backward compatibility requires that userspace observe
659 how the kernel module parses received packets. This is only
660 possible in a straightforward way when userspace adds kernel flows
661 in reaction to received packets.
663 For more relevant information on the architecture of Open vSwitch,
664 please read "The Design and Implementation of Open vSwitch",
665 published in USENIX NSDI 2015.
671 ### Q: I just upgraded and I see a performance drop. Why?
673 A: The OVS kernel datapath may have been updated to a newer version than
674 the OVS userspace components. Sometimes new versions of OVS kernel
675 module add functionality that is backwards compatible with older
676 userspace components but may cause a drop in performance with them.
677 Especially, if a kernel module from OVS 2.1 or newer is paired with
678 OVS userspace 1.10 or older, there will be a performance drop for
681 Updating the OVS userspace components to the latest released
682 version should fix the performance degradation.
684 To get the best possible performance and functionality, it is
685 recommended to pair the same versions of the kernel module and OVS
689 Configuration Problems
690 ----------------------
692 ### Q: I created a bridge and added my Ethernet port to it, using commands
696 ovs-vsctl add-port br0 eth0
698 and as soon as I ran the "add-port" command I lost all connectivity
701 A: A physical Ethernet device that is part of an Open vSwitch bridge
702 should not have an IP address. If one does, then that IP address
703 will not be fully functional.
705 You can restore functionality by moving the IP address to an Open
706 vSwitch "internal" device, such as the network device named after
707 the bridge itself. For example, assuming that eth0's IP address is
708 192.168.128.5, you could run the commands below to fix up the
711 ifconfig eth0 0.0.0.0
712 ifconfig br0 192.168.128.5
714 (If your only connection to the machine running OVS is through the
715 IP address in question, then you would want to run all of these
716 commands on a single command line, or put them into a script.) If
717 there were any additional routes assigned to eth0, then you would
718 also want to use commands to adjust these routes to go through br0.
720 If you use DHCP to obtain an IP address, then you should kill the
721 DHCP client that was listening on the physical Ethernet interface
722 (e.g. eth0) and start one listening on the internal interface
723 (e.g. br0). You might still need to manually clear the IP address
724 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
726 There is no compelling reason why Open vSwitch must work this way.
727 However, this is the way that the Linux kernel bridge module has
728 always worked, so it's a model that those accustomed to Linux
729 bridging are already used to. Also, the model that most people
730 expect is not implementable without kernel changes on all the
731 versions of Linux that Open vSwitch supports.
733 By the way, this issue is not specific to physical Ethernet
734 devices. It applies to all network devices except Open vSwitch
737 ### Q: I created a bridge and added a couple of Ethernet ports to it,
738 ### using commands like these:
741 ovs-vsctl add-port br0 eth0
742 ovs-vsctl add-port br0 eth1
744 and now my network seems to have melted: connectivity is unreliable
745 (even connectivity that doesn't go through Open vSwitch), all the
746 LEDs on my physical switches are blinking, wireshark shows
747 duplicated packets, and CPU usage is very high.
749 A: More than likely, you've looped your network. Probably, eth0 and
750 eth1 are connected to the same physical Ethernet switch. This
751 yields a scenario where OVS receives a broadcast packet on eth0 and
752 sends it out on eth1, then the physical switch connected to eth1
753 sends the packet back on eth0, and so on forever. More complicated
754 scenarios, involving a loop through multiple switches, are possible
757 The solution depends on what you are trying to do:
759 - If you added eth0 and eth1 to get higher bandwidth or higher
760 reliability between OVS and your physical Ethernet switch,
761 use a bond. The following commands create br0 and then add
762 eth0 and eth1 as a bond:
765 ovs-vsctl add-bond br0 bond0 eth0 eth1
767 Bonds have tons of configuration options. Please read the
768 documentation on the Port table in ovs-vswitchd.conf.db(5)
771 Configuration for DPDK-enabled interfaces is slightly less
772 straightforward: see [INSTALL.DPDK.md].
774 - Perhaps you don't actually need eth0 and eth1 to be on the
775 same bridge. For example, if you simply want to be able to
776 connect each of them to virtual machines, then you can put
777 each of them on a bridge of its own:
780 ovs-vsctl add-port br0 eth0
783 ovs-vsctl add-port br1 eth1
785 and then connect VMs to br0 and br1. (A potential
786 disadvantage is that traffic cannot directly pass between br0
787 and br1. Instead, it will go out eth0 and come back in eth1,
790 - If you have a redundant or complex network topology and you
791 want to prevent loops, turn on spanning tree protocol (STP).
792 The following commands create br0, enable STP, and add eth0
793 and eth1 to the bridge. The order is important because you
794 don't want have to have a loop in your network even
798 ovs-vsctl set bridge br0 stp_enable=true
799 ovs-vsctl add-port br0 eth0
800 ovs-vsctl add-port br0 eth1
802 The Open vSwitch implementation of STP is not well tested.
803 Please report any bugs you observe, but if you'd rather avoid
804 acting as a beta tester then another option might be your
807 ### Q: I can't seem to use Open vSwitch in a wireless network.
809 A: Wireless base stations generally only allow packets with the source
810 MAC address of NIC that completed the initial handshake.
811 Therefore, without MAC rewriting, only a single device can
812 communicate over a single wireless link.
814 This isn't specific to Open vSwitch, it's enforced by the access
815 point, so the same problems will show up with the Linux bridge or
816 any other way to do bridging.
818 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
820 A: PPP most commonly carries IP packets, but Open vSwitch works only
821 with Ethernet frames. The correct way to interface PPP to an
822 Ethernet network is usually to use routing instead of switching.
824 ### Q: Is there any documentation on the database tables and fields?
826 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
828 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
829 I only see a datapath called "ovs-system". How can I see datapath
830 information about a particular bridge?
832 A: In version 1.9.0, OVS switched to using a single datapath that is
833 shared by all bridges of that type. The "ovs-appctl dpif/*"
834 commands provide similar functionality that is scoped by the bridge.
836 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
837 see the port in the datapath?
839 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
840 cannot be loaded at the same time. It is likely that on your system the
841 Linux GRE module is already loaded and blocking OVS (to confirm, check
842 dmesg for errors regarding GRE registration). To fix this, unload all
843 GRE modules that appear in lsmod as well as the OVS kernel module. You
844 can then reload the OVS module following the directions in
845 [INSTALL.md], which will ensure that dependencies are satisfied.
847 ### Q: Open vSwitch does not seem to obey my packet filter rules.
849 A: It depends on mechanisms and configurations you want to use.
851 You cannot usefully use typical packet filters, like iptables, on
852 physical Ethernet ports that you add to an Open vSwitch bridge.
853 This is because Open vSwitch captures packets from the interface at
854 a layer lower below where typical packet-filter implementations
855 install their hooks. (This actually applies to any interface of
856 type "system" that you might add to an Open vSwitch bridge.)
858 You can usefully use typical packet filters on Open vSwitch
859 internal ports as they are mostly ordinary interfaces from the point
860 of view of packet filters.
862 For example, suppose you create a bridge br0 and add Ethernet port
863 eth0 to it. Then you can usefully add iptables rules to affect the
864 internal interface br0, but not the physical interface eth0. (br0
865 is also where you would add an IP address, as discussed elsewhere
868 For simple filtering rules, it might be possible to achieve similar
869 results by installing appropriate OpenFlow flows instead.
871 If the use of a particular packet filter setup is essential, Open
872 vSwitch might not be the best choice for you. On Linux, you might
873 want to consider using the Linux Bridge. (This is the only choice if
874 you want to use ebtables rules.) On NetBSD, you might want to
875 consider using the bridge(4) with BRIDGE_IPF option.
877 ### Q: It seems that Open vSwitch does nothing when I removed a port and
878 then immediately put it back. For example, consider that p1 is
879 a port of type=internal:
881 ovs-vsctl del-port br0 p1 -- \
883 set interface p1 type=internal
885 A: It's an expected behaviour.
887 If del-port and add-port happen in a single OVSDB transaction as
888 your example, Open vSwitch always "skips" the intermediate steps.
889 Even if they are done in multiple transactions, it's still allowed
890 for Open vSwitch to skip the intermediate steps and just implement
891 the overall effect. In both cases, your example would be turned
894 If you want to make Open vSwitch actually destroy and then re-create
895 the port for some side effects like resetting kernel setting for the
896 corresponding interface, you need to separate operations into multiple
897 OVSDB transactions and ensure that at least the first one does not have
898 --no-wait. In the following example, the first ovs-vsctl will block
899 until Open vSwitch reloads the new configuration and removes the port:
901 ovs-vsctl del-port br0 p1
902 ovs-vsctl add-port br0 p1 -- \
903 set interface p1 type=internal
905 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
906 it takes too long (minutes or hours) to do it with ovs-vsctl. How
909 A: If you add them one at a time with ovs-vsctl, it can take a long
910 time to add thousands of ports to an Open vSwitch bridge. This is
911 because every invocation of ovs-vsctl first reads the current
912 configuration from OVSDB. As the number of ports grows, this
913 starts to take an appreciable amount of time, and when it is
914 repeated thousands of times the total time becomes significant.
916 The solution is to add the ports in one invocation of ovs-vsctl (or
917 a small number of them). For example, using bash:
920 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
923 takes seconds, not minutes or hours, in the OVS sandbox environment.
925 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
926 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
929 A: Open vSwitch wasn't able to create the bridge. Check the
930 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
931 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
933 In general, the Open vSwitch database reflects the desired
934 configuration state. ovs-vswitchd monitors the database and, when
935 it changes, reconfigures the system to reflect the new desired
936 state. This normally happens very quickly. Thus, a discrepancy
937 between the database and the actual state indicates that
938 ovs-vswitchd could not implement the configuration, and so one
939 should check the log to find out why. (Another possible cause is
940 that ovs-vswitchd is not running. This will make "ovs-vsctl"
941 commands hang, if they change the configuration, unless one
942 specifies "--no-wait".)
944 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
945 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
946 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
949 A: Open vSwitch wasn't able to create the port. Check the
950 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
951 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
952 see the previous question for more information.
954 You may want to upgrade to Open vSwitch 2.3 (or later), in which
955 ovs-vsctl will immediately report when there is an issue creating a
958 ### Q: I created a tap device tap0, configured an IP address on it, and
959 added it to a bridge, like this:
962 ifconfig tap0 192.168.0.123
964 ovs-vsctl add-port br0 tap0
966 I expected that I could then use this IP address to contact other
967 hosts on the network, but it doesn't work. Why not?
969 A: The short answer is that this is a misuse of a "tap" device. Use
970 an "internal" device implemented by Open vSwitch, which works
971 differently and is designed for this use. To solve this problem
972 with an internal device, instead run:
975 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
976 ifconfig int0 192.168.0.123
978 Even more simply, you can take advantage of the internal port that
979 every bridge has under the name of the bridge:
982 ifconfig br0 192.168.0.123
984 In more detail, a "tap" device is an interface between the Linux
985 (or *BSD) network stack and a user program that opens it as a
986 socket. When the "tap" device transmits a packet, it appears in
987 the socket opened by the userspace program. Conversely, when the
988 userspace program writes to the "tap" socket, the kernel TCP/IP
989 stack processes the packet as if it had been received by the "tap"
992 Consider the configuration above. Given this configuration, if you
993 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
994 routing stack will transmit an ARP on the tap0 device. Open
995 vSwitch userspace treats "tap" devices just like any other network
996 device; that is, it doesn't open them as "tap" sockets. That means
997 that the ARP packet will simply get dropped.
999 You might wonder why the Open vSwitch kernel module doesn't
1000 intercept the ARP packet and bridge it. After all, Open vSwitch
1001 intercepts packets on other devices. The answer is that Open
1002 vSwitch only intercepts *received* packets, but this is a packet
1003 being transmitted. The same thing happens for all other types of
1004 network devices, except for Open vSwitch "internal" ports. If you,
1005 for example, add a physical Ethernet port to an OVS bridge,
1006 configure an IP address on a physical Ethernet port, and then issue
1007 a "ping" to an address in that subnet, the same thing happens: an
1008 ARP gets transmitted on the physical Ethernet port and Open vSwitch
1009 never sees it. (You should not do that, as documented at the
1010 beginning of this section.)
1012 It can make sense to add a "tap" device to an Open vSwitch bridge,
1013 if some userspace program (other than Open vSwitch) has opened the
1014 tap socket. This is the case, for example, if the "tap" device was
1015 created by KVM (or QEMU) to simulate a virtual NIC. In such a
1016 case, when OVS bridges a packet to the "tap" device, the kernel
1017 forwards that packet to KVM in userspace, which passes it along to
1018 the VM, and in the other direction, when the VM sends a packet, KVM
1019 writes it to the "tap" socket, which causes OVS to receive it and
1020 bridge it to the other OVS ports. Please note that in such a case
1021 no IP address is configured on the "tap" device (there is normally
1022 an IP address configured in the virtual NIC inside the VM, but this
1023 is not visible to the host Linux kernel or to Open vSwitch).
1025 There is one special case in which Open vSwitch does directly read
1026 and write "tap" sockets. This is an implementation detail of the
1027 Open vSwitch userspace switch, which implements its "internal"
1028 ports as Linux (or *BSD) "tap" sockets. In such a userspace
1029 switch, OVS receives packets sent on the "tap" device used to
1030 implement an "internal" port by reading the associated "tap"
1031 socket, and bridges them to the rest of the switch. In the other
1032 direction, OVS transmits packets bridged to the "internal" port by
1033 writing them to the "tap" socket, causing them to be processed by
1034 the kernel TCP/IP stack as if they had been received on the "tap"
1035 device. Users should not need to be concerned with this
1036 implementation detail.
1038 Open vSwitch has a network device type called "tap". This is
1039 intended only for implementing "internal" ports in the OVS
1040 userspace switch and should not be used otherwise. In particular,
1041 users should not configure KVM "tap" devices as type "tap" (use
1042 type "system", the default, instead).
1045 Quality of Service (QoS)
1046 ------------------------
1048 ### Q: Does OVS support Quality of Service (QoS)?
1050 A: Yes. For traffic that egresses from a switch, OVS supports traffic
1051 shaping; for traffic that ingresses into a switch, OVS support
1052 policing. Policing is a simple form of quality-of-service that
1053 simply drops packets received in excess of the configured rate. Due
1054 to its simplicity, policing is usually less accurate and less
1055 effective than egress traffic shaping, which queues packets.
1057 Keep in mind that ingress and egress are from the perspective of the
1058 switch. That means that egress shaping limits the rate at which
1059 traffic is allowed to transmit from a physical interface, but not the
1060 rate at which traffic will be received on a virtual machine's VIF.
1061 For ingress policing, the behavior is the opposite.
1063 ### Q: How do I configure egress traffic shaping?
1065 A: Suppose that you want to set up bridge br0 connected to physical
1066 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
1067 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
1068 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
1069 could configure the bridge this way:
1073 add-port br0 eth0 -- \
1074 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
1075 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
1076 set port eth0 qos=@newqos -- \
1077 --id=@newqos create qos type=linux-htb \
1078 other-config:max-rate=1000000000 \
1079 queues:123=@vif10queue \
1080 queues:234=@vif20queue -- \
1081 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
1082 --id=@vif20queue create queue other-config:max-rate=20000000
1084 At this point, bridge br0 is configured with the ports and eth0 is
1085 configured with the queues that you need for QoS, but nothing is
1086 actually directing packets from vif1.0 or vif2.0 to the queues that
1087 we have set up for them. That means that all of the packets to
1088 eth0 are going to the "default queue", which is not what we want.
1090 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
1091 queues reserved for them:
1093 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
1094 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
1096 Each of the above flows matches on the input port, sets up the
1097 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
1098 executes the "normal" action, which performs the same switching
1099 that Open vSwitch would have done without any OpenFlow flows being
1100 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1101 numbers 5 and 6, respectively, because we set their ofport_request
1102 columns above. If we had not done that, then we would have needed
1103 to find out their port numbers before setting up these flows.)
1105 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1108 By the way, if you delete the bridge created by the above commands,
1111 ovs-vsctl del-br br0
1113 then that will leave one unreferenced QoS record and two
1114 unreferenced Queue records in the Open vSwich database. One way to
1115 clear them out, assuming you don't have other QoS or Queue records
1116 that you want to keep, is:
1118 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1120 If you do want to keep some QoS or Queue records, or the Open
1121 vSwitch you are using is older than version 1.8 (which added the
1122 --all option), then you will have to destroy QoS and Queue records
1125 ### Q: How do I configure ingress policing?
1127 A: A policing policy can be configured on an interface to drop packets
1128 that arrive at a higher rate than the configured value. For example,
1129 the following commands will rate-limit traffic that vif1.0 may
1132 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1133 ovs-vsctl set interface vif1.0 ingress_policing_burst=8000
1135 Traffic policing can interact poorly with some network protocols and
1136 can have surprising results. The "Ingress Policing" section of
1137 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1139 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1140 adding records to the QoS and Queue table, but the results aren't
1143 A: Did you install OpenFlow flows that use your queues? This is the
1144 primary way to tell Open vSwitch which queues you want to use. If
1145 you don't do this, then the default queue will be used, which will
1146 probably not have the effect you want.
1148 Refer to the previous question for an example.
1150 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1151 doesn't yet support. How do I do that?
1153 A: Open vSwitch does not implement QoS itself. Instead, it can
1154 configure some, but not all, of the QoS features built into the
1155 Linux kernel. If you need some QoS feature that OVS cannot
1156 configure itself, then the first step is to figure out whether
1157 Linux QoS supports that feature. If it does, then you can submit a
1158 patch to support Open vSwitch configuration for that feature, or
1159 you can use "tc" directly to configure the feature in Linux. (If
1160 Linux QoS doesn't support the feature you want, then first you have
1161 to add that support to Linux.)
1163 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1164 working as well as I expect.
1166 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1169 - Open vSwitch configures a subset of Linux kernel QoS
1170 features, according to what is in OVSDB. It is possible that
1171 this code has bugs. If you believe that this is so, then you
1172 can configure the Linux traffic control (QoS) stack directly
1173 with the "tc" program. If you get better results that way,
1174 you can send a detailed bug report to bugs@openvswitch.org.
1176 It is certain that Open vSwitch cannot configure every Linux
1177 kernel QoS feature. If you need some feature that OVS cannot
1178 configure, then you can also use "tc" directly (or add that
1181 - The Open vSwitch implementation of OpenFlow allows flows to
1182 be directed to particular queues. This is pretty simple and
1183 unlikely to have serious bugs at this point.
1185 However, most problems with QoS on Linux are not bugs in Open
1186 vSwitch at all. They tend to be either configuration errors
1187 (please see the earlier questions in this section) or issues with
1188 the traffic control (QoS) stack in Linux. The Open vSwitch
1189 developers are not experts on Linux traffic control. We suggest
1190 that, if you believe you are encountering a problem with Linux
1191 traffic control, that you consult the tc manpages (e.g. tc(8),
1192 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1193 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1195 ### Q: Does Open vSwitch support OpenFlow meters?
1197 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1198 OpenFlow meters. There is no implementation of meters in the Open
1199 vSwitch software switch (neither the kernel-based nor userspace
1206 ### Q: What's a VLAN?
1208 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1209 partition a single switch into multiple switches. Suppose, for
1210 example, that you have two groups of machines, group A and group B.
1211 You want the machines in group A to be able to talk to each other,
1212 and you want the machine in group B to be able to talk to each
1213 other, but you don't want the machines in group A to be able to
1214 talk to the machines in group B. You can do this with two
1215 switches, by plugging the machines in group A into one switch and
1216 the machines in group B into the other switch.
1218 If you only have one switch, then you can use VLANs to do the same
1219 thing, by configuring the ports for machines in group A as VLAN
1220 "access ports" for one VLAN and the ports for group B as "access
1221 ports" for a different VLAN. The switch will only forward packets
1222 between ports that are assigned to the same VLAN, so this
1223 effectively subdivides your single switch into two independent
1224 switches, one for each group of machines.
1226 So far we haven't said anything about VLAN headers. With access
1227 ports, like we've described so far, no VLAN header is present in
1228 the Ethernet frame. This means that the machines (or switches)
1229 connected to access ports need not be aware that VLANs are
1230 involved, just like in the case where we use two different physical
1233 Now suppose that you have a whole bunch of switches in your
1234 network, instead of just one, and that some machines in group A are
1235 connected directly to both switches 1 and 2. To allow these
1236 machines to talk to each other, you could add an access port for
1237 group A's VLAN to switch 1 and another to switch 2, and then
1238 connect an Ethernet cable between those ports. That works fine,
1239 but it doesn't scale well as the number of switches and the number
1240 of VLANs increases, because you use up a lot of valuable switch
1241 ports just connecting together your VLANs.
1243 This is where VLAN headers come in. Instead of using one cable and
1244 two ports per VLAN to connect a pair of switches, we configure a
1245 port on each switch as a VLAN "trunk port". Packets sent and
1246 received on a trunk port carry a VLAN header that says what VLAN
1247 the packet belongs to, so that only two ports total are required to
1248 connect the switches, regardless of the number of VLANs in use.
1249 Normally, only switches (either physical or virtual) are connected
1250 to a trunk port, not individual hosts, because individual hosts
1251 don't expect to see a VLAN header in the traffic that they receive.
1253 None of the above discussion says anything about particular VLAN
1254 numbers. This is because VLAN numbers are completely arbitrary.
1255 One must only ensure that a given VLAN is numbered consistently
1256 throughout a network and that different VLANs are given different
1257 numbers. (That said, VLAN 0 is usually synonymous with a packet
1258 that has no VLAN header, and VLAN 4095 is reserved.)
1260 ### Q: VLANs don't work.
1262 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1263 bugs. If you are having problems with VLANs that you suspect to be
1264 driver related, then you have several options:
1266 - Upgrade to Linux 3.3 or later.
1268 - Build and install a fixed version of the particular driver
1269 that is causing trouble, if one is available.
1271 - Use a NIC whose driver does not have VLAN problems.
1273 - Use "VLAN splinters", a feature in Open vSwitch 1.4 upto 2.5
1274 that works around bugs in kernel drivers. To enable VLAN
1275 splinters on interface eth0, use the command:
1277 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1279 For VLAN splinters to be effective, Open vSwitch must know
1280 which VLANs are in use. See the "VLAN splinters" section in
1281 the Interface table in ovs-vswitchd.conf.db(5) for details on
1282 how Open vSwitch infers in-use VLANs.
1284 VLAN splinters increase memory use and reduce performance, so
1285 use them only if needed.
1287 - Apply the "vlan workaround" patch from the XenServer kernel
1288 patch queue, build Open vSwitch against this patched kernel,
1289 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1290 workaround for each interface whose driver is buggy.
1292 (This is a nontrivial exercise, so this option is included
1293 only for completeness.)
1295 It is not always easy to tell whether a Linux kernel driver has
1296 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1297 can help you test. See their manpages for details. Of the two
1298 utilities, ovs-test(8) is newer and more thorough, but
1299 ovs-vlan-test(8) may be easier to use.
1301 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1303 A: Do you have VLANs enabled on the physical switch that OVS is
1304 attached to? Make sure that the port is configured to trunk the
1305 VLAN or VLANs that you are using with OVS.
1307 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1308 and to its destination host, but OVS seems to drop incoming return
1311 A: It's possible that you have the VLAN configured on your physical
1312 switch as the "native" VLAN. In this mode, the switch treats
1313 incoming packets either tagged with the native VLAN or untagged as
1314 part of the native VLAN. It may also send outgoing packets in the
1315 native VLAN without a VLAN tag.
1317 If this is the case, you have two choices:
1319 - Change the physical switch port configuration to tag packets
1320 it forwards to OVS with the native VLAN instead of forwarding
1323 - Change the OVS configuration for the physical port to a
1324 native VLAN mode. For example, the following sets up a
1325 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1327 ovs-vsctl add-br br0
1328 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1330 In this situation, "native-untagged" mode will probably work
1331 equally well. Refer to the documentation for the Port table
1332 in ovs-vswitchd.conf.db(5) for more information.
1334 ### Q: I added a pair of VMs on different VLANs, like this:
1336 ovs-vsctl add-br br0
1337 ovs-vsctl add-port br0 eth0
1338 ovs-vsctl add-port br0 tap0 tag=9
1339 ovs-vsctl add-port br0 tap1 tag=10
1341 but the VMs can't access each other, the external network, or the
1344 A: It is to be expected that the VMs can't access each other. VLANs
1345 are a means to partition a network. When you configured tap0 and
1346 tap1 as access ports for different VLANs, you indicated that they
1347 should be isolated from each other.
1349 As for the external network and the Internet, it seems likely that
1350 the machines you are trying to access are not on VLAN 9 (or 10) and
1351 that the Internet is not available on VLAN 9 (or 10).
1353 ### Q: I added a pair of VMs on the same VLAN, like this:
1355 ovs-vsctl add-br br0
1356 ovs-vsctl add-port br0 eth0
1357 ovs-vsctl add-port br0 tap0 tag=9
1358 ovs-vsctl add-port br0 tap1 tag=9
1360 The VMs can access each other, but not the external network or the
1363 A: It seems likely that the machines you are trying to access in the
1364 external network are not on VLAN 9 and that the Internet is not
1365 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1366 trunk VLAN on the upstream switch port to which eth0 is connected.
1368 ### Q: Can I configure an IP address on a VLAN?
1370 A: Yes. Use an "internal port" configured as an access port. For
1371 example, the following configures IP address 192.168.0.7 on VLAN 9.
1372 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1373 they have an 802.1Q header with VLAN 9. Conversely, traffic
1374 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1377 ovs-vsctl add-br br0
1378 ovs-vsctl add-port br0 eth0
1379 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1380 ifconfig vlan9 192.168.0.7
1382 See also the following question.
1384 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1387 ovs-vsctl add-br br0
1388 ovs-vsctl add-port br0 eth0
1389 ifconfig br0 192.168.0.5
1390 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1391 ifconfig vlan9 192.168.0.9
1393 but other hosts that are only on VLAN 0 can reach the IP address
1394 configured on VLAN 9. What's going on?
1396 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1397 approaches to IP address handling in Internet hosts:
1399 - In the "Strong ES Model", where an ES is a host ("End
1400 System"), an IP address is primarily associated with a
1401 particular interface. The host discards packets that arrive
1402 on interface A if they are destined for an IP address that is
1403 configured on interface B. The host never sends packets from
1404 interface A using a source address configured on interface B.
1406 - In the "Weak ES Model", an IP address is primarily associated
1407 with a host. The host accepts packets that arrive on any
1408 interface if they are destined for any of the host's IP
1409 addresses, even if the address is configured on some
1410 interface other than the one on which it arrived. The host
1411 does not restrict itself to sending packets from an IP
1412 address associated with the originating interface.
1414 Linux uses the weak ES model. That means that when packets
1415 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1416 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1417 address, even though they were not received on vlan9, the network
1420 To simulate the strong ES model on Linux, one may add iptables rule
1421 to filter packets based on source and destination address and
1422 adjust ARP configuration with sysctls.
1424 BSD uses the strong ES model.
1426 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1428 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1429 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1430 implementation of the OpenFlow "normal switching" action. By
1431 default, when Open vSwitch isn't connected to a controller and
1432 nothing has been manually configured in the flow table, all traffic
1433 goes through the "normal switching" action. But, if you set up
1434 OpenFlow flows on your own, through a controller or using ovs-ofctl
1435 or through other means, then you have to implement VLAN handling
1438 You can use "normal switching" as a component of your OpenFlow
1439 actions, e.g. by putting "normal" into the lists of actions on
1440 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1441 controller. In situations where this is not suitable, you can
1442 implement VLAN handling yourself, e.g.:
1444 - If a packet comes in on an access port, and the flow table
1445 needs to send it out on a trunk port, then the flow can add
1446 the appropriate VLAN tag with the "mod_vlan_vid" action.
1448 - If a packet comes in on a trunk port, and the flow table
1449 needs to send it out on an access port, then the flow can
1450 strip the VLAN tag with the "strip_vlan" action.
1452 ### Q: I configured ports on a bridge as access ports with different VLAN
1455 ovs-vsctl add-br br0
1456 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1457 ovs-vsctl add-port br0 eth0
1458 ovs-vsctl add-port br0 tap0 tag=9
1459 ovs-vsctl add-port br0 tap1 tag=10
1461 but the VMs running behind tap0 and tap1 can still communicate,
1462 that is, they are not isolated from each other even though they are
1465 A: Do you have a controller configured on br0 (as the commands above
1466 do)? If so, then this is a variant on the previous question, "My
1467 OpenFlow controller doesn't see the VLANs that I expect," and you
1468 can refer to the answer there for more information.
1470 ### Q: How MAC learning works with VLANs?
1472 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1473 OFPP_NORMAL action. I.e. it logically has separate learning tables
1480 ### Q: What's a VXLAN?
1482 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1483 to solve the scaling challenges of VLAN networks in a multi-tenant
1484 environment. VXLAN is an overlay network which transports an L2 network
1485 over an existing L3 network. For more information on VXLAN, please see
1488 http://tools.ietf.org/html/rfc7348
1490 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1492 A: Open vSwitch currently supports the framing format for packets on the
1493 wire. There is currently no support for the multicast aspects of VXLAN.
1494 To get around the lack of multicast support, it is possible to
1495 pre-provision MAC to IP address mappings either manually or from a
1498 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1501 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1502 is 4789. However, it is possible to configure the destination UDP port
1503 manually on a per-VXLAN tunnel basis. An example of this configuration is
1506 ovs-vsctl add-br br0
1507 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1508 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1509 options:dst_port=8472
1512 Using OpenFlow (Manually or Via Controller)
1513 -------------------------------------------
1515 ### Q: What versions of OpenFlow does Open vSwitch support?
1517 A: The following table lists the versions of OpenFlow supported by
1518 each version of Open vSwitch:
1520 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5 OF1.6
1521 ###============ ===== ===== ===== ===== ===== ===== =====
1522 1.9 and earlier yes --- --- --- --- --- ---
1523 1.10 yes --- [*] [*] --- --- ---
1524 1.11 yes --- [*] [*] --- --- ---
1525 2.0 yes [*] [*] [*] --- --- ---
1526 2.1 yes [*] [*] [*] --- --- ---
1527 2.2 yes [*] [*] [*] [%] [*] ---
1528 2.3 yes yes yes yes [*] [*] ---
1529 2.4 yes yes yes yes [*] [*] ---
1530 2.5 yes yes yes yes [*] [*] [*]
1532 [*] Supported, with one or more missing features.
1533 [%] Experimental, unsafe implementation.
1535 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1536 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1537 1.2, and 1.3 must be enabled manually in ovs-vswitchd.
1539 Some versions of OpenFlow are supported with missing features and
1540 therefore not enabled by default: OpenFlow 1.4 and 1.5, in Open
1541 vSwitch 2.3 and later, as well as OpenFlow 1.6 in Open vSwitch 2.5
1542 and later. Also, the OpenFlow 1.6 specification is still under
1543 development and thus subject to change.
1545 In any case, the user may override the default:
1547 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1549 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1551 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1553 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1555 - To enable only OpenFlow 1.0 on bridge br0:
1557 ovs-vsctl set bridge br0 protocols=OpenFlow10
1559 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1560 default. Use the -O option to enable support for later versions of
1561 OpenFlow in ovs-ofctl. For example:
1563 ovs-ofctl -O OpenFlow13 dump-flows br0
1565 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1566 1.4 that could cause crashes. We don't recommend enabling it.)
1568 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1569 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1570 1.5 is solidly implemented, Open vSwitch will enable those version
1573 ### Q: Does Open vSwitch support MPLS?
1575 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1576 these versions can match on MPLS Ethernet types, but they cannot
1577 match, push, or pop MPLS labels, nor can they look past MPLS labels
1578 into the encapsulated packet.
1580 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1581 for MPLS. With the userspace datapath only, these versions can
1582 match, push, or pop a single MPLS label, but they still cannot look
1583 past MPLS labels (even after popping them) into the encapsulated
1584 packet. Kernel datapath support is unchanged from earlier
1587 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1588 label and look past the MPLS label into the encapsulated packet.
1589 Both userspace and kernel datapaths will be supported, but MPLS
1590 processing always happens in userspace either way, so kernel
1591 datapath performance will be disappointing.
1593 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1594 labels and look past the MPLS label into the encapsulated packet.
1595 It will have kernel support for MPLS, yielding improved
1598 ### Q: I'm getting "error type 45250 code 0". What's that?
1600 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1601 vSwitch uses this extension when it must report an error to an
1602 OpenFlow controller but no standard OpenFlow error code is
1605 Open vSwitch logs the errors that it sends to controllers, so the
1606 easiest thing to do is probably to look at the ovs-vswitchd log to
1607 find out what the error was.
1609 If you want to dissect the extended error message yourself, the
1610 format is documented in include/openflow/nicira-ext.h in the Open
1611 vSwitch source distribution. The extended error codes are
1612 documented in include/openvswitch/ofp-errors.h.
1614 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1615 doesn't actually appear through the OpenFlow connection, even
1616 though I know that it's going through.
1617 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1618 to apply to certain traffic, especially traffic between OVS and
1619 the controller itself.
1621 A: By default, Open vSwitch assumes that OpenFlow controllers are
1622 connected "in-band", that is, that the controllers are actually
1623 part of the network that is being controlled. In in-band mode,
1624 Open vSwitch sets up special "hidden" flows to make sure that
1625 traffic can make it back and forth between OVS and the controllers.
1626 These hidden flows are higher priority than any flows that can be
1627 set up through OpenFlow, and they are not visible through normal
1628 OpenFlow flow table dumps.
1630 Usually, the hidden flows are desirable and helpful, but
1631 occasionally they can cause unexpected behavior. You can view the
1632 full OpenFlow flow table, including hidden flows, on bridge br0
1635 ovs-appctl bridge/dump-flows br0
1637 to help you debug. The hidden flows are those with priorities
1638 greater than 65535 (the maximum priority that can be set with
1641 The DESIGN file at the top level of the Open vSwitch source
1642 distribution describes the in-band model in detail.
1644 If your controllers are not actually in-band (e.g. they are on
1645 localhost via 127.0.0.1, or on a separate network), then you should
1646 configure your controllers in "out-of-band" mode. If you have one
1647 controller on bridge br0, then you can configure out-of-band mode
1650 ovs-vsctl set controller br0 connection-mode=out-of-band
1652 ### Q: I configured all my controllers for out-of-band control mode but
1653 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1655 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1656 set-manager"). By default, Open vSwitch assumes that managers need
1657 in-band rules set up on every bridge. You can disable these rules
1660 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1662 This actually disables in-band control entirely for the bridge, as
1663 if all the bridge's controllers were configured for out-of-band
1666 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1668 A: See answer under "VLANs", above.
1670 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1671 but I got a funny message like this:
1673 ofp_util|INFO|normalization changed ofp_match, details:
1674 ofp_util|INFO| pre: nw_dst=192.168.0.1
1677 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1678 match had disappeared, so that the flow ends up matching every
1681 A: The term "normalization" in the log message means that a flow
1682 cannot match on an L3 field without saying what L3 protocol is in
1683 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1684 so the L3 field match was dropped.
1686 In this case, the L3 protocol could be IP or ARP. A correct
1687 command for each possibility is, respectively:
1689 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1693 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1695 Similarly, a flow cannot match on an L4 field without saying what
1696 L4 protocol is in use. For example, the flow match "tp_src=1234"
1697 is, by itself, meaningless and will be ignored. Instead, to match
1698 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1699 source port 1234, write "udp,tp_src=1234".
1701 ### Q: How can I figure out the OpenFlow port number for a given port?
1703 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1704 respond with an OFPT_FEATURES_REPLY that, among other information,
1705 includes a mapping between OpenFlow port names and numbers. From a
1706 command prompt, "ovs-ofctl show br0" makes such a request and
1707 prints the response for switch br0.
1709 The Interface table in the Open vSwitch database also maps OpenFlow
1710 port names to numbers. To print the OpenFlow port number
1711 associated with interface eth0, run:
1713 ovs-vsctl get Interface eth0 ofport
1715 You can print the entire mapping with:
1717 ovs-vsctl -- --columns=name,ofport list Interface
1719 but the output mixes together interfaces from all bridges in the
1720 database, so it may be confusing if more than one bridge exists.
1722 In the Open vSwitch database, ofport value -1 means that the
1723 interface could not be created due to an error. (The Open vSwitch
1724 log should indicate the reason.) ofport value [] (the empty set)
1725 means that the interface hasn't been created yet. The latter is
1726 normally an intermittent condition (unless ovs-vswitchd is not
1729 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1730 run "ovs-dpctl dump-flows" I don't see them.
1732 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1733 won't display the information that you want. You want to use
1734 "ovs-ofctl dump-flows" instead.
1736 ### Q: It looks like each of the interfaces in my bonded port shows up
1737 as an individual OpenFlow port. Is that right?
1739 A: Yes, Open vSwitch makes individual bond interfaces visible as
1740 OpenFlow ports, rather than the bond as a whole. The interfaces
1741 are treated together as a bond for only a few purposes:
1743 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1744 controller is not configured, this happens implicitly to
1747 - Mirrors configured for output to a bonded port.
1749 It would make a lot of sense for Open vSwitch to present a bond as
1750 a single OpenFlow port. If you want to contribute an
1751 implementation of such a feature, please bring it up on the Open
1752 vSwitch development mailing list at dev@openvswitch.org.
1754 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1755 multiple hosts, and other components. The behavior isn't what I
1758 A: To debug network behavior problems, trace the path of a packet,
1759 hop-by-hop, from its origin in one host to a remote host. If
1760 that's correct, then trace the path of the response packet back to
1763 The open source tool called "plotnetcfg" can help to understand the
1764 relationship between the networking devices on a single host.
1766 Usually a simple ICMP echo request and reply ("ping") packet is
1767 good enough. Start by initiating an ongoing "ping" from the origin
1768 host to a remote host. If you are tracking down a connectivity
1769 problem, the "ping" will not display any successful output, but
1770 packets are still being sent. (In this case the packets being sent
1771 are likely ARP rather than ICMP.)
1773 Tools available for tracing include the following:
1775 - "tcpdump" and "wireshark" for observing hops across network
1776 devices, such as Open vSwitch internal devices and physical
1779 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1780 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1781 These tools allow one to observe the actions being taken on
1782 packets in ongoing flows.
1784 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1785 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1786 documentation, and "Why are there so many different ways to
1787 dump flows?" above for some background.
1789 - "ovs-appctl ofproto/trace" to observe the logic behind how
1790 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1791 documentation. You can out more details about a given flow
1792 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1793 a flow from the output into an "ovs-appctl ofproto/trace"
1796 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1797 observe what goes on at these physical hops.
1799 Starting at the origin of a given packet, observe the packet at
1800 each hop in turn. For example, in one plausible scenario, you
1803 1. "tcpdump" the "eth" interface through which an ARP egresses
1804 a VM, from inside the VM.
1806 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1807 ingresses the host machine.
1809 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1810 the host interface through which the ARP egresses the
1811 physical machine. You may need to use "ovs-dpctl show" to
1812 interpret the port numbers. If the output seems surprising,
1813 you can use "ovs-appctl ofproto/trace" to observe details of
1814 how ovs-vswitchd determined the actions in the "ovs-dpctl
1817 4. "tcpdump" the "eth" interface through which the ARP egresses
1818 the physical machine.
1820 5. "tcpdump" the "eth" interface through which the ARP
1821 ingresses the physical machine, at the remote host that
1824 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1825 remote host that receives the ARP and observe the VM "vif"
1826 or "tap" interface to which the flow is directed. Again,
1827 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1829 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1832 8. "tcpdump" the "eth" interface through which the ARP
1833 ingresses a VM, from inside the VM.
1835 It is likely that during one of these steps you will figure out the
1836 problem. If not, then follow the ARP reply back to the origin, in
1839 ### Q: How do I make a flow drop packets?
1841 A: To drop a packet is to receive it without forwarding it. OpenFlow
1842 explicitly specifies forwarding actions. Thus, a flow with an
1843 empty set of actions does not forward packets anywhere, causing
1844 them to be dropped. You can specify an empty set of actions with
1845 "actions=" on the ovs-ofctl command line. For example:
1847 ovs-ofctl add-flow br0 priority=65535,actions=
1849 would cause every packet entering switch br0 to be dropped.
1851 You can write "drop" explicitly if you like. The effect is the
1852 same. Thus, the following command also causes every packet
1853 entering switch br0 to be dropped:
1855 ovs-ofctl add-flow br0 priority=65535,actions=drop
1857 "drop" is not an action, either in OpenFlow or Open vSwitch.
1858 Rather, it is only a way to say that there are no actions.
1860 ### Q: I added a flow to send packets out the ingress port, like this:
1862 ovs-ofctl add-flow br0 in_port=2,actions=2
1864 but OVS drops the packets instead.
1866 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1867 out its ingress port. The rationale is that dropping these packets
1868 makes it harder to loop the network. Sometimes this behavior can
1869 even be convenient, e.g. it is often the desired behavior in a flow
1870 that forwards a packet to several ports ("floods" the packet).
1872 Sometimes one really needs to send a packet out its ingress port
1873 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1874 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1876 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1878 This also works in some circumstances where the flow doesn't match
1879 on the input port. For example, if you know that your switch has
1880 five ports numbered 2 through 6, then the following will send every
1881 received packet out every port, even its ingress port:
1883 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1887 ovs-ofctl add-flow br0 actions=all,in_port
1889 Sometimes, in complicated flow tables with multiple levels of
1890 "resubmit" actions, a flow needs to output to a particular port
1891 that may or may not be the ingress port. It's difficult to take
1892 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1893 provides, as an OpenFlow extension, the ability to modify the
1894 in_port field. Whatever value is currently in the in_port field is
1895 the port to which outputs will be dropped, as well as the
1896 destination for OFPP_IN_PORT. This means that the following will
1897 reliably output to port 2 or to ports 2 through 6, respectively:
1899 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1900 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1902 If the input port is important, then one may save and restore it on
1905 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1906 load:0->NXM_OF_IN_PORT[],\
1908 pop:NXM_OF_IN_PORT[]
1910 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1911 on port 2. I set up flows to forward only traffic destined to the
1912 other host and drop other traffic, like this:
1914 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1915 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1916 priority=0,actions=drop
1918 But it doesn't work--I don't get any connectivity when I do this.
1921 A: These flows drop the ARP packets that IP hosts use to establish IP
1922 connectivity over Ethernet. To solve the problem, add flows to
1923 allow ARP to pass between the hosts:
1925 priority=5,in_port=1,arp,actions=2
1926 priority=5,in_port=2,arp,actions=1
1928 This issue can manifest other ways, too. The following flows that
1929 match on Ethernet addresses instead of IP addresses will also drop
1930 ARP packets, because ARP requests are broadcast instead of being
1931 directed to a specific host:
1933 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1934 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1935 priority=0,actions=drop
1937 The solution already described above will also work in this case.
1938 It may be better to add flows to allow all multicast and broadcast
1941 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1942 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1944 ### Q: My bridge disconnects from my controller on add-port/del-port.
1946 A: Reconfiguring your bridge can change your bridge's datapath-id because
1947 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1948 In that case, Open vSwitch disconnects from controllers because there's
1949 no graceful way to notify controllers about the change of datapath-id.
1951 To avoid the behaviour, you can configure datapath-id manually.
1953 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1955 ### Q: My controller is getting errors about "buffers". What's going on?
1957 A: When a switch sends a packet to an OpenFlow controller using a
1958 "packet-in" message, it can also keep a copy of that packet in a
1959 "buffer", identified by a 32-bit integer "buffer_id". There are
1960 two advantages to buffering. First, when the controller wants to
1961 tell the switch to do something with the buffered packet (with a
1962 "packet-out" OpenFlow request), it does not need to send another
1963 copy of the packet back across the OpenFlow connection, which
1964 reduces the bandwidth cost of the connection and improves latency.
1965 This enables the second advantage: the switch can optionally send
1966 only the first part of the packet to the controller (assuming that
1967 the switch only needs to look at the first few bytes of the
1968 packet), further reducing bandwidth and improving latency.
1970 However, buffering introduces some issues of its own. First, any
1971 switch has limited resources, so if the controller does not use a
1972 buffered packet, the switch has to decide how long to keep it
1973 buffered. When many packets are sent to a controller and buffered,
1974 Open vSwitch can discard buffered packets that the controller has
1975 not used after as little as 5 seconds. This means that
1976 controllers, if they make use of packet buffering, should use the
1977 buffered packets promptly. (This includes sending a "packet-out"
1978 with no actions if the controller does not want to do anything with
1979 a buffered packet, to clear the packet buffer and effectively
1982 Second, packet buffers are one-time-use, meaning that a controller
1983 cannot use a single packet buffer in two or more "packet-out"
1984 commands. Open vSwitch will respond with an error to the second
1985 and subsequent "packet-out"s in such a case.
1987 Finally, a common error early in controller development is to try
1988 to use buffer_id 0 in a "packet-out" message as if 0 represented
1989 "no buffered packet". This is incorrect usage: the buffer_id with
1990 this meaning is actually 0xffffffff.
1992 ovs-vswitchd(8) describes some details of Open vSwitch packet
1993 buffering that the OpenFlow specification requires implementations
1996 ### Q: How does OVS divide flows among buckets in an OpenFlow "select" group?
1998 A: In Open vSwitch 2.3 and earlier, Open vSwitch used the destination
1999 Ethernet address to choose a bucket in a select group.
2001 Open vSwitch 2.4 and later by default hashes the source and
2002 destination Ethernet address, VLAN ID, Ethernet type, IPv4/v6
2003 source and destination address and protocol, and for TCP and SCTP
2004 only, the source and destination ports. The hash is "symmetric",
2005 meaning that exchanging source and destination addresses does not
2006 change the bucket selection.
2008 Select groups in Open vSwitch 2.4 and later can be configured to
2009 use a different hash function, using a Netronome extension to the
2010 OpenFlow 1.5+ group_mod message. For more information, see
2011 Documentation/group-selection-method-property.txt in the Open
2012 vSwitch source tree. (OpenFlow 1.5 support in Open vSwitch is still
2015 ### Q: I added a flow to accept packets on VLAN 123 and output them on
2018 ovs-ofctl add-flow br0 dl_vlan=123,actions=output:1,mod_vlan_vid:456
2020 but the packets are actually being output in VLAN 123. Why?
2022 A: OpenFlow actions are executed in the order specified. Thus, the
2023 actions above first output the packet, then change its VLAN. Since
2024 the output occurs before changing the VLAN, the change in VLAN will
2025 have no visible effect.
2027 To solve this and similar problems, order actions so that changes
2028 to headers happen before output, e.g.:
2030 ovs-ofctl add-flow br0 dl_vlan=123,actions=mod_vlan_vid:456,output:1
2036 ### Q: How do I implement a new OpenFlow message?
2038 A: Add your new message to "enum ofpraw" and "enum ofptype" in
2039 lib/ofp-msgs.h, following the existing pattern. Then recompile and
2040 fix all of the new warnings, implementing new functionality for the
2041 new message as needed. (If you configure with --enable-Werror, as
2042 described in [INSTALL.md], then it is impossible to miss any warnings.)
2044 If you need to add an OpenFlow vendor extension message for a
2045 vendor that doesn't yet have any extension messages, then you will
2046 also need to edit build-aux/extract-ofp-msgs.
2048 ### Q: How do I add support for a new field or header?
2050 A: Add new members for your field to "struct flow" in lib/flow.h, and
2051 add new enumerations for your new field to "enum mf_field_id" in
2052 lib/meta-flow.h, following the existing pattern. Also, add support
2053 to miniflow_extract() in lib/flow.c for extracting your new field
2054 from a packet into struct miniflow, and to nx_put_raw() in
2055 lib/nx-match.c to output your new field in OXM matches. Then
2056 recompile and fix all of the new warnings, implementing new
2057 functionality for the new field or header as needed. (If you
2058 configure with --enable-Werror, as described in [INSTALL.md], then
2059 it is impossible to miss any warnings.)
2061 If you want kernel datapath support for your new field, you also
2062 need to modify the kernel module for the operating systems you are
2063 interested in. This isn't mandatory, since fields understood only
2064 by userspace work too (with a performance penalty), so it's
2065 reasonable to start development without it. If you implement
2066 kernel module support for Linux, then the Linux kernel "netdev"
2067 mailing list is the place to submit that support first; please read
2068 up on the Linux kernel development process separately. The Windows
2069 datapath kernel module support, on the other hand, is maintained
2070 within the OVS tree, so patches for that can go directly to
2073 ### Q: How do I add support for a new OpenFlow action?
2075 A: Add your new action to "enum ofp_raw_action_type" in
2076 lib/ofp-actions.c, following the existing pattern. Then recompile
2077 and fix all of the new warnings, implementing new functionality for
2078 the new action as needed. (If you configure with --enable-Werror,
2079 as described in [INSTALL.md], then it is impossible to miss any
2082 If you need to add an OpenFlow vendor extension action for a vendor
2083 that doesn't yet have any extension actions, then you will also
2084 need to edit build-aux/extract-ofp-actions.
2090 bugs@openvswitch.org
2091 http://openvswitch.org/
2093 [PORTING.md]:PORTING.md
2094 [WHY-OVS.md]:WHY-OVS.md
2095 [INSTALL.md]:INSTALL.md
2096 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
2097 [INSTALL.DPDK.md]:INSTALL.DPDK.md
projects / cascardo / ovs.git / blob