1 Frequently Asked Questions
2 ==========================
4 Open vSwitch <http://openvswitch.org>
9 ### Q: What is Open vSwitch?
11 A: Open vSwitch is a production quality open source software switch
12 designed to be used as a vswitch in virtualized server
13 environments. A vswitch forwards traffic between different VMs on
14 the same physical host and also forwards traffic between VMs and
15 the physical network. Open vSwitch supports standard management
16 interfaces (e.g. sFlow, NetFlow, IPFIX, RSPAN, CLI), and is open to
17 programmatic extension and control using OpenFlow and the OVSDB
20 Open vSwitch as designed to be compatible with modern switching
21 chipsets. This means that it can be ported to existing high-fanout
22 switches allowing the same flexible control of the physical
23 infrastructure as the virtual infrastructure. It also means that
24 Open vSwitch will be able to take advantage of on-NIC switching
25 chipsets as their functionality matures.
27 ### Q: What virtualization platforms can use Open vSwitch?
29 A: Open vSwitch can currently run on any Linux-based virtualization
30 platform (kernel 3.10 and newer), including: KVM, VirtualBox, Xen,
31 Xen Cloud Platform, XenServer. As of Linux 3.3 it is part of the
32 mainline kernel. The bulk of the code is written in platform-
33 independent C and is easily ported to other environments. We welcome
34 inquires about integrating Open vSwitch with other virtualization
37 ### Q: How can I try Open vSwitch?
39 A: The Open vSwitch source code can be built on a Linux system. You can
40 build and experiment with Open vSwitch on any Linux machine.
41 Packages for various Linux distributions are available on many
42 platforms, including: Debian, Ubuntu, Fedora.
44 You may also download and run a virtualization platform that already
45 has Open vSwitch integrated. For example, download a recent ISO for
46 XenServer or Xen Cloud Platform. Be aware that the version
47 integrated with a particular platform may not be the most recent Open
50 ### Q: Does Open vSwitch only work on Linux?
52 A: No, Open vSwitch has been ported to a number of different operating
53 systems and hardware platforms. Most of the development work occurs
54 on Linux, but the code should be portable to any POSIX system. We've
55 seen Open vSwitch ported to a number of different platforms,
56 including FreeBSD, Windows, and even non-POSIX embedded systems.
58 By definition, the Open vSwitch Linux kernel module only works on
59 Linux and will provide the highest performance. However, a userspace
60 datapath is available that should be very portable.
62 ### Q: What's involved with porting Open vSwitch to a new platform or switching ASIC?
64 A: The [PORTING.md] document describes how one would go about
65 porting Open vSwitch to a new operating system or hardware platform.
67 ### Q: Why would I use Open vSwitch instead of the Linux bridge?
69 A: Open vSwitch is specially designed to make it easier to manage VM
70 network configuration and monitor state spread across many physical
71 hosts in dynamic virtualized environments. Please see
72 [WHY-OVS.md] for a more detailed description of how Open vSwitch
73 relates to the Linux Bridge.
75 ### Q: How is Open vSwitch related to distributed virtual switches like the VMware vNetwork distributed switch or the Cisco Nexus 1000V?
77 A: Distributed vswitch applications (e.g., VMware vNetwork distributed
78 switch, Cisco Nexus 1000V) provide a centralized way to configure and
79 monitor the network state of VMs that are spread across many physical
80 hosts. Open vSwitch is not a distributed vswitch itself, rather it
81 runs on each physical host and supports remote management in a way
82 that makes it easier for developers of virtualization/cloud
83 management platforms to offer distributed vswitch capabilities.
85 To aid in distribution, Open vSwitch provides two open protocols that
86 are specially designed for remote management in virtualized network
87 environments: OpenFlow, which exposes flow-based forwarding state,
88 and the OVSDB management protocol, which exposes switch port state.
89 In addition to the switch implementation itself, Open vSwitch
90 includes tools (ovs-ofctl, ovs-vsctl) that developers can script and
91 extend to provide distributed vswitch capabilities that are closely
92 integrated with their virtualization management platform.
94 ### Q: Why doesn't Open vSwitch support distribution?
96 A: Open vSwitch is intended to be a useful component for building
97 flexible network infrastructure. There are many different approaches
98 to distribution which balance trade-offs between simplicity,
99 scalability, hardware compatibility, convergence times, logical
100 forwarding model, etc. The goal of Open vSwitch is to be able to
101 support all as a primitive building block rather than choose a
102 particular point in the distributed design space.
104 ### Q: How can I contribute to the Open vSwitch Community?
106 A: You can start by joining the mailing lists and helping to answer
107 questions. You can also suggest improvements to documentation. If
108 you have a feature or bug you would like to work on, send a mail to
109 one of the mailing lists:
111 http://openvswitch.org/mlists/
113 ### Q: Why can I no longer connect to my OpenFlow controller or OVSDB manager?
115 A: Starting in OVS 2.4, we switched the default ports to the
116 IANA-specified port numbers for OpenFlow (6633->6653) and OVSDB
117 (6632->6640). We recommend using these port numbers, but if you
118 cannot, all the programs allow overriding the default port. See the
119 appropriate man page.
125 ### Q: What does it mean for an Open vSwitch release to be LTS (long-term support)?
127 A: All official releases have been through a comprehensive testing
128 process and are suitable for production use. Planned releases will
129 occur several times a year. If a significant bug is identified in an
130 LTS release, we will provide an updated release that includes the
131 fix. Releases that are not LTS may not be fixed and may just be
132 supplanted by the next major release. The current LTS release is
135 ### Q: What Linux kernel versions does each Open vSwitch release work with?
137 A: The following table lists the Linux kernel versions against which the
138 given versions of the Open vSwitch kernel module will successfully
139 build. The Linux kernel versions are upstream kernel versions, so
140 Linux kernels modified from the upstream sources may not build in
141 some cases even if they are based on a supported version. This is
142 most notably true of Red Hat Enterprise Linux (RHEL) kernels, which
143 are extensively modified from upstream.
145 | Open vSwitch | Linux kernel
146 |:------------:|:-------------:
147 | 1.4.x | 2.6.18 to 3.2
148 | 1.5.x | 2.6.18 to 3.2
149 | 1.6.x | 2.6.18 to 3.2
150 | 1.7.x | 2.6.18 to 3.3
151 | 1.8.x | 2.6.18 to 3.4
152 | 1.9.x | 2.6.18 to 3.8
153 | 1.10.x | 2.6.18 to 3.8
154 | 1.11.x | 2.6.18 to 3.8
155 | 2.0.x | 2.6.32 to 3.10
156 | 2.1.x | 2.6.32 to 3.11
157 | 2.3.x | 2.6.32 to 3.14
158 | 2.4.x | 2.6.32 to 4.0
159 | 2.5.x | 2.6.32 to 4.3
160 | 2.6.x | 3.10 to 4.3
162 Open vSwitch userspace should also work with the Linux kernel module
163 built into Linux 3.3 and later.
165 Open vSwitch userspace is not sensitive to the Linux kernel version.
166 It should build against almost any kernel, certainly against 2.6.32
169 ### Q: Are all features available with all datapaths?
171 A: Open vSwitch supports different datapaths on different platforms. Each
172 datapath has a different feature set: the following tables try to summarize
177 * *Linux upstream*: The datapath implemented by the kernel module shipped
178 with Linux upstream. Since features have been gradually
179 introduced into the kernel, the table mentions the first
180 Linux release whose OVS module supports the feature.
182 * *Linux OVS tree*: The datapath implemented by the Linux kernel module
183 distributed with the OVS source tree. Some features of
184 this module rely on functionality not available in older
185 kernels: in this case the minumum Linux version (against
186 which the feature can be compiled) is listed.
188 * *Userspace*: Also known as DPDK, dpif-netdev or dummy datapath. It is the
189 only datapath that works on NetBSD and FreeBSD.
191 * *Hyper-V*: Also known as the Windows datapath.
193 The following table lists the datapath supported features from
194 an Open vSwitch user's perspective.
196 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
197 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
198 Connection tracking | 4.3 | 3.10 | NO | NO |
199 Tunnel - LISP | NO | YES | NO | NO |
200 Tunnel - STT | NO | 3.5 | NO | YES |
201 Tunnel - GRE | 3.11 | YES | YES | YES |
202 Tunnel - VXLAN | 3.12 | YES | YES | YES |
203 Tunnel - Geneve | 3.18 | YES | YES | NO |
204 QoS - Policing | YES | YES | NO | NO |
205 QoS - Shaping | YES | YES | NO | NO |
206 sFlow | YES | YES | YES | NO |
207 Set action | YES | YES | YES | PARTIAL |
208 NIC Bonding | YES | YES | YES | NO |
209 Multiple VTEPs | YES | YES | YES | NO |
212 * Only a limited set of flow fields is modifiable via the set action by the
214 * The Hyper-V datapath only supports one physical NIC per datapath. This is
215 why bonding is not supported.
216 * The Hyper-V datapath can have at most one IP address configured as a
219 The following table lists features that do not *directly* impact an
220 Open vSwitch user, e.g. because their absence can be hidden by the ofproto
221 layer (usually this comes with a performance penalty).
223 Feature | Linux upstream | Linux OVS tree | Userspace | Hyper-V |
224 ----------------------|:--------------:|:--------------:|:---------:|:-------:|
225 SCTP flows | 3.12 | YES | YES | YES |
226 MPLS | 3.19 | YES | YES | YES |
227 UFID | 4.0 | YES | YES | NO |
228 Megaflows | 3.12 | YES | YES | NO |
229 Masked set action | 4.0 | YES | YES | NO |
230 Recirculation | 3.19 | YES | YES | NO |
231 TCP flags matching | 3.13 | YES | YES | NO |
232 Validate flow actions | YES | YES | N/A | NO |
233 Multiple datapaths | YES | YES | YES | NO |
234 Tunnel TSO - STT | N/A | YES | NO | YES |
236 ### Q: I get an error like this when I configure Open vSwitch:
238 configure: error: Linux kernel in <dir> is version <x>, but
239 version newer than <y> is not supported (please refer to the
244 A: You have the following options:
246 - Use the Linux kernel module supplied with the kernel that you are
247 using. (See also the following FAQ.)
249 - If there is a newer released version of Open vSwitch, consider
250 building that one, because it may support the kernel that you are
251 building against. (To find out, consult the table in the
254 - The Open vSwitch "master" branch may support the kernel that you
255 are using, so consider building the kernel module from "master".
257 All versions of Open vSwitch userspace are compatible with all
258 versions of the Open vSwitch kernel module, so you do not have to
259 use the kernel module from one source along with the userspace
260 programs from the same source.
262 ### Q: What features are not available in the Open vSwitch kernel datapath that ships as part of the upstream Linux kernel?
264 A: The kernel module in upstream Linux does not include support for
265 LISP. Work is in progress to add support for LISP to the upstream
266 Linux version of the Open vSwitch kernel module. For now, if you
267 need this feature, use the kernel module from the Open vSwitch
268 distribution instead of the upstream Linux kernel module.
270 Certain features require kernel support to function or to have
271 reasonable performance. If the ovs-vswitchd log file indicates that
272 a feature is not supported, consider upgrading to a newer upstream
273 Linux release or using the kernel module paired with the userspace
276 ### Q: Why do tunnels not work when using a kernel module other than the one packaged with Open vSwitch?
278 A: Support for tunnels was added to the upstream Linux kernel module
279 after the rest of Open vSwitch. As a result, some kernels may contain
280 support for Open vSwitch but not tunnels. The minimum kernel version
281 that supports each tunnel protocol is:
283 | Protocol | Linux Kernel
284 |:--------:|:-------------:
288 | LISP | <not upstream>
289 | STT | <not upstream>
291 If you are using a version of the kernel that is older than the one
292 listed above, it is still possible to use that tunnel protocol. However,
293 you must compile and install the kernel module included with the Open
294 vSwitch distribution rather than the one on your machine. If problems
295 persist after doing this, check to make sure that the module that is
296 loaded is the one you expect.
298 ### Q: Why are UDP tunnel checksums not computed for VXLAN or Geneve?
300 A: Generating outer UDP checksums requires kernel support that was not
301 part of the initial implementation of these protocols. If using the
302 upstream Linux Open vSwitch module, you must use kernel 4.0 or
303 newer. The out-of-tree modules from Open vSwitch release 2.4 and later
304 support UDP checksums.
306 ### Q: What features are not available when using the userspace datapath?
308 A: Tunnel virtual ports are not supported, as described in the
309 previous answer. It is also not possible to use queue-related
310 actions. On Linux kernels before 2.6.39, maximum-sized VLAN packets
311 may not be transmitted.
313 ### Q: What Linux kernel versions does IPFIX flow monitoring work with?
315 A: IPFIX flow monitoring requires the Linux kernel module from Linux
316 3.10 or later, or the out-of-tree module from Open vSwitch version
319 ### Q: Should userspace or kernel be upgraded first to minimize downtime?
321 In general, the Open vSwitch userspace should be used with the
322 kernel version included in the same release or with the version
323 from upstream Linux. However, when upgrading between two releases
324 of Open vSwitch it is best to migrate userspace first to reduce
325 the possibility of incompatibilities.
327 ### Q: What happened to the bridge compatibility feature?
329 A: Bridge compatibility was a feature of Open vSwitch 1.9 and earlier.
330 When it was enabled, Open vSwitch imitated the interface of the
331 Linux kernel "bridge" module. This allowed users to drop Open
332 vSwitch into environments designed to use the Linux kernel bridge
333 module without adapting the environment to use Open vSwitch.
335 Open vSwitch 1.10 and later do not support bridge compatibility.
336 The feature was dropped because version 1.10 adopted a new internal
337 architecture that made bridge compatibility difficult to maintain.
338 Now that many environments use OVS directly, it would be rarely
341 To use bridge compatibility, install OVS 1.9 or earlier, including
342 the accompanying kernel modules (both the main and bridge
343 compatibility modules), following the instructions that come with
344 the release. Be sure to start the ovs-brcompatd daemon.
350 ### Q: I thought Open vSwitch was a virtual Ethernet switch, but the documentation keeps talking about bridges. What's a bridge?
352 A: In networking, the terms "bridge" and "switch" are synonyms. Open
353 vSwitch implements an Ethernet switch, which means that it is also
356 ### Q: What's a VLAN?
358 A: See the "VLAN" section below.
364 ### Q: How do I configure a port as an access port?
366 A: Add "tag=VLAN" to your "ovs-vsctl add-port" command. For example,
367 the following commands configure br0 with eth0 as a trunk port (the
368 default) and tap0 as an access port for VLAN 9:
371 ovs-vsctl add-port br0 eth0
372 ovs-vsctl add-port br0 tap0 tag=9
374 If you want to configure an already added port as an access port,
375 use "ovs-vsctl set", e.g.:
377 ovs-vsctl set port tap0 tag=9
379 ### Q: How do I configure a port as a SPAN port, that is, enable mirroring of all traffic to that port?
381 A: The following commands configure br0 with eth0 and tap0 as trunk
382 ports. All traffic coming in or going out on eth0 or tap0 is also
383 mirrored to tap1; any traffic arriving on tap1 is dropped:
386 ovs-vsctl add-port br0 eth0
387 ovs-vsctl add-port br0 tap0
388 ovs-vsctl add-port br0 tap1 \
389 -- --id=@p get port tap1 \
390 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
391 -- set bridge br0 mirrors=@m
393 To later disable mirroring, run:
395 ovs-vsctl clear bridge br0 mirrors
397 ### Q: Does Open vSwitch support configuring a port in promiscuous mode?
399 A: Yes. How you configure it depends on what you mean by "promiscuous
402 - Conventionally, "promiscuous mode" is a feature of a network
403 interface card. Ordinarily, a NIC passes to the CPU only the
404 packets actually destined to its host machine. It discards
405 the rest to avoid wasting memory and CPU cycles. When
406 promiscuous mode is enabled, however, it passes every packet
407 to the CPU. On an old-style shared-media or hub-based
408 network, this allows the host to spy on all packets on the
409 network. But in the switched networks that are almost
410 everywhere these days, promiscuous mode doesn't have much
411 effect, because few packets not destined to a host are
412 delivered to the host's NIC.
414 This form of promiscuous mode is configured in the guest OS of
415 the VMs on your bridge, e.g. with "ifconfig".
417 - The VMware vSwitch uses a different definition of "promiscuous
418 mode". When you configure promiscuous mode on a VMware vNIC,
419 the vSwitch sends a copy of every packet received by the
420 vSwitch to that vNIC. That has a much bigger effect than just
421 enabling promiscuous mode in a guest OS. Rather than getting
422 a few stray packets for which the switch does not yet know the
423 correct destination, the vNIC gets every packet. The effect
424 is similar to replacing the vSwitch by a virtual hub.
426 This "promiscuous mode" is what switches normally call "port
427 mirroring" or "SPAN". For information on how to configure
428 SPAN, see "How do I configure a port as a SPAN port, that is,
429 enable mirroring of all traffic to that port?"
431 ### Q: How do I configure a DPDK port as an access port?
433 A: Firstly, you must have a DPDK-enabled version of Open vSwitch.
435 If your version is DPDK-enabled it will support the --dpdk
436 argument on the command line and will display lines with
437 "EAL:..." during startup when --dpdk is supplied.
439 Secondly, when adding a DPDK port, unlike a system port, the
440 type for the interface must be specified. For example;
443 ovs-vsctl add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk
445 Finally, it is required that DPDK port names begin with 'dpdk'.
447 See [INSTALL.DPDK.md] for more information on enabling and using DPDK with
450 ### Q: How do I configure a VLAN as an RSPAN VLAN, that is, enable mirroring of all traffic to that VLAN?
452 A: The following commands configure br0 with eth0 as a trunk port and
453 tap0 as an access port for VLAN 10. All traffic coming in or going
454 out on tap0, as well as traffic coming in or going out on eth0 in
455 VLAN 10, is also mirrored to VLAN 15 on eth0. The original tag for
456 VLAN 10, in cases where one is present, is dropped as part of
460 ovs-vsctl add-port br0 eth0
461 ovs-vsctl add-port br0 tap0 tag=10
463 -- --id=@m create mirror name=m0 select-all=true select-vlan=10 \
465 -- set bridge br0 mirrors=@m
467 To later disable mirroring, run:
469 ovs-vsctl clear bridge br0 mirrors
471 Mirroring to a VLAN can disrupt a network that contains unmanaged
472 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
473 GRE tunnel has fewer caveats than mirroring to a VLAN and should
474 generally be preferred.
476 ### Q: Can I mirror more than one input VLAN to an RSPAN VLAN?
478 A: Yes, but mirroring to a VLAN strips the original VLAN tag in favor
479 of the specified output-vlan. This loss of information may make
480 the mirrored traffic too hard to interpret.
482 To mirror multiple VLANs, use the commands above, but specify a
483 comma-separated list of VLANs as the value for select-vlan. To
484 mirror every VLAN, use the commands above, but omit select-vlan and
487 When a packet arrives on a VLAN that is used as a mirror output
488 VLAN, the mirror is disregarded. Instead, in standalone mode, OVS
489 floods the packet across all the ports for which the mirror output
490 VLAN is configured. (If an OpenFlow controller is in use, then it
491 can override this behavior through the flow table.) If OVS is used
492 as an intermediate switch, rather than an edge switch, this ensures
493 that the RSPAN traffic is distributed through the network.
495 Mirroring to a VLAN can disrupt a network that contains unmanaged
496 switches. See ovs-vswitchd.conf.db(5) for details. Mirroring to a
497 GRE tunnel has fewer caveats than mirroring to a VLAN and should
498 generally be preferred.
500 ### Q: How do I configure mirroring of all traffic to a GRE tunnel?
502 A: The following commands configure br0 with eth0 and tap0 as trunk
503 ports. All traffic coming in or going out on eth0 or tap0 is also
504 mirrored to gre0, a GRE tunnel to the remote host 192.168.1.10; any
505 traffic arriving on gre0 is dropped:
508 ovs-vsctl add-port br0 eth0
509 ovs-vsctl add-port br0 tap0
510 ovs-vsctl add-port br0 gre0 \
511 -- set interface gre0 type=gre options:remote_ip=192.168.1.10 \
512 -- --id=@p get port gre0 \
513 -- --id=@m create mirror name=m0 select-all=true output-port=@p \
514 -- set bridge br0 mirrors=@m
516 To later disable mirroring and destroy the GRE tunnel:
518 ovs-vsctl clear bridge br0 mirrors
519 ovs-vsctl del-port br0 gre0
521 ### Q: Does Open vSwitch support ERSPAN?
523 A: No. ERSPAN is an undocumented proprietary protocol. As an
524 alternative, Open vSwitch supports mirroring to a GRE tunnel (see
527 ### Q: How do I connect two bridges?
529 A: First, why do you want to do this? Two connected bridges are not
530 much different from a single bridge, so you might as well just have
531 a single bridge with all your ports on it.
533 If you still want to connect two bridges, you can use a pair of
534 patch ports. The following example creates bridges br0 and br1,
535 adds eth0 and tap0 to br0, adds tap1 to br1, and then connects br0
536 and br1 with a pair of patch ports.
539 ovs-vsctl add-port br0 eth0
540 ovs-vsctl add-port br0 tap0
542 ovs-vsctl add-port br1 tap1
544 -- add-port br0 patch0 \
545 -- set interface patch0 type=patch options:peer=patch1 \
546 -- add-port br1 patch1 \
547 -- set interface patch1 type=patch options:peer=patch0
549 Bridges connected with patch ports are much like a single bridge.
550 For instance, if the example above also added eth1 to br1, and both
551 eth0 and eth1 happened to be connected to the same next-hop switch,
552 then you could loop your network just as you would if you added
553 eth0 and eth1 to the same bridge (see the "Configuration Problems"
554 section below for more information).
556 If you are using Open vSwitch 1.9 or an earlier version, then you
557 need to be using the kernel module bundled with Open vSwitch rather
558 than the one that is integrated into Linux 3.3 and later, because
559 Open vSwitch 1.9 and earlier versions need kernel support for patch
560 ports. This also means that in Open vSwitch 1.9 and earlier, patch
561 ports will not work with the userspace datapath, only with the
564 ### Q: How do I configure a bridge without an OpenFlow local port? (Local port in the sense of OFPP_LOCAL)
566 A: Open vSwitch does not support such a configuration.
567 Bridges always have their local ports.
570 Implementation Details
571 ----------------------
573 ### Q: I hear OVS has a couple of kinds of flows. Can you tell me about them?
575 A: Open vSwitch uses different kinds of flows for different purposes:
577 - OpenFlow flows are the most important kind of flow. OpenFlow
578 controllers use these flows to define a switch's policy.
579 OpenFlow flows support wildcards, priorities, and multiple
582 When in-band control is in use, Open vSwitch sets up a few
583 "hidden" flows, with priority higher than a controller or the
584 user can configure, that are not visible via OpenFlow. (See
585 the "Controller" section of the FAQ for more information
588 - The Open vSwitch software switch implementation uses a second
589 kind of flow internally. These flows, called "datapath" or
590 "kernel" flows, do not support priorities and comprise only a
591 single table, which makes them suitable for caching. (Like
592 OpenFlow flows, datapath flows do support wildcarding, in Open
593 vSwitch 1.11 and later.) OpenFlow flows and datapath flows
594 also support different actions and number ports differently.
596 Datapath flows are an implementation detail that is subject to
597 change in future versions of Open vSwitch. Even with the
598 current version of Open vSwitch, hardware switch
599 implementations do not necessarily use this architecture.
601 Users and controllers directly control only the OpenFlow flow
602 table. Open vSwitch manages the datapath flow table itself, so
603 users should not normally be concerned with it.
605 ### Q: Why are there so many different ways to dump flows?
607 A: Open vSwitch has two kinds of flows (see the previous question), so
608 it has commands with different purposes for dumping each kind of
611 - `ovs-ofctl dump-flows <br>` dumps OpenFlow flows, excluding
612 hidden flows. This is the most commonly useful form of flow
613 dump. (Unlike the other commands, this should work with any
614 OpenFlow switch, not just Open vSwitch.)
616 - `ovs-appctl bridge/dump-flows <br>` dumps OpenFlow flows,
617 including hidden flows. This is occasionally useful for
618 troubleshooting suspected issues with in-band control.
620 - `ovs-dpctl dump-flows [dp]` dumps the datapath flow table
621 entries for a Linux kernel-based datapath. In Open vSwitch
622 1.10 and later, ovs-vswitchd merges multiple switches into a
623 single datapath, so it will show all the flows on all your
624 kernel-based switches. This command can occasionally be
625 useful for debugging.
627 - `ovs-appctl dpif/dump-flows <br>`, new in Open vSwitch 1.10,
628 dumps datapath flows for only the specified bridge, regardless
631 ### Q: How does multicast snooping works with VLANs?
633 A: Open vSwitch maintains snooping tables for each VLAN.
635 ### Q: Can OVS populate the kernel flow table in advance instead of in reaction to packets?
637 A: No. There are several reasons:
639 - Kernel flows are not as sophisticated as OpenFlow flows, which
640 means that some OpenFlow policies could require a large number of
641 kernel flows. The "conjunctive match" feature is an extreme
642 example: the number of kernel flows it requires is the product of
643 the number of flows in each dimension.
645 - With multiple OpenFlow flow tables and simple sets of actions, the
646 number of kernel flows required can be as large as the product of
647 the number of flows in each dimension. With more sophisticated
648 actions, the number of kernel flows could be even larger.
650 - Open vSwitch is designed so that any version of OVS userspace
651 interoperates with any version of the OVS kernel module. This
652 forward and backward compatibility requires that userspace observe
653 how the kernel module parses received packets. This is only
654 possible in a straightforward way when userspace adds kernel flows
655 in reaction to received packets.
657 For more relevant information on the architecture of Open vSwitch,
658 please read "The Design and Implementation of Open vSwitch",
659 published in USENIX NSDI 2015.
665 ### Q: I just upgraded and I see a performance drop. Why?
667 A: The OVS kernel datapath may have been updated to a newer version than
668 the OVS userspace components. Sometimes new versions of OVS kernel
669 module add functionality that is backwards compatible with older
670 userspace components but may cause a drop in performance with them.
671 Especially, if a kernel module from OVS 2.1 or newer is paired with
672 OVS userspace 1.10 or older, there will be a performance drop for
675 Updating the OVS userspace components to the latest released
676 version should fix the performance degradation.
678 To get the best possible performance and functionality, it is
679 recommended to pair the same versions of the kernel module and OVS
683 Configuration Problems
684 ----------------------
686 ### Q: I created a bridge and added my Ethernet port to it, using commands
690 ovs-vsctl add-port br0 eth0
692 and as soon as I ran the "add-port" command I lost all connectivity
695 A: A physical Ethernet device that is part of an Open vSwitch bridge
696 should not have an IP address. If one does, then that IP address
697 will not be fully functional.
699 You can restore functionality by moving the IP address to an Open
700 vSwitch "internal" device, such as the network device named after
701 the bridge itself. For example, assuming that eth0's IP address is
702 192.168.128.5, you could run the commands below to fix up the
705 ifconfig eth0 0.0.0.0
706 ifconfig br0 192.168.128.5
708 (If your only connection to the machine running OVS is through the
709 IP address in question, then you would want to run all of these
710 commands on a single command line, or put them into a script.) If
711 there were any additional routes assigned to eth0, then you would
712 also want to use commands to adjust these routes to go through br0.
714 If you use DHCP to obtain an IP address, then you should kill the
715 DHCP client that was listening on the physical Ethernet interface
716 (e.g. eth0) and start one listening on the internal interface
717 (e.g. br0). You might still need to manually clear the IP address
718 from the physical interface (e.g. with "ifconfig eth0 0.0.0.0").
720 There is no compelling reason why Open vSwitch must work this way.
721 However, this is the way that the Linux kernel bridge module has
722 always worked, so it's a model that those accustomed to Linux
723 bridging are already used to. Also, the model that most people
724 expect is not implementable without kernel changes on all the
725 versions of Linux that Open vSwitch supports.
727 By the way, this issue is not specific to physical Ethernet
728 devices. It applies to all network devices except Open vSwitch
731 ### Q: I created a bridge and added a couple of Ethernet ports to it,
732 ### using commands like these:
735 ovs-vsctl add-port br0 eth0
736 ovs-vsctl add-port br0 eth1
738 and now my network seems to have melted: connectivity is unreliable
739 (even connectivity that doesn't go through Open vSwitch), all the
740 LEDs on my physical switches are blinking, wireshark shows
741 duplicated packets, and CPU usage is very high.
743 A: More than likely, you've looped your network. Probably, eth0 and
744 eth1 are connected to the same physical Ethernet switch. This
745 yields a scenario where OVS receives a broadcast packet on eth0 and
746 sends it out on eth1, then the physical switch connected to eth1
747 sends the packet back on eth0, and so on forever. More complicated
748 scenarios, involving a loop through multiple switches, are possible
751 The solution depends on what you are trying to do:
753 - If you added eth0 and eth1 to get higher bandwidth or higher
754 reliability between OVS and your physical Ethernet switch,
755 use a bond. The following commands create br0 and then add
756 eth0 and eth1 as a bond:
759 ovs-vsctl add-bond br0 bond0 eth0 eth1
761 Bonds have tons of configuration options. Please read the
762 documentation on the Port table in ovs-vswitchd.conf.db(5)
765 Configuration for DPDK-enabled interfaces is slightly less
766 straightforward: see [INSTALL.DPDK.md].
768 - Perhaps you don't actually need eth0 and eth1 to be on the
769 same bridge. For example, if you simply want to be able to
770 connect each of them to virtual machines, then you can put
771 each of them on a bridge of its own:
774 ovs-vsctl add-port br0 eth0
777 ovs-vsctl add-port br1 eth1
779 and then connect VMs to br0 and br1. (A potential
780 disadvantage is that traffic cannot directly pass between br0
781 and br1. Instead, it will go out eth0 and come back in eth1,
784 - If you have a redundant or complex network topology and you
785 want to prevent loops, turn on spanning tree protocol (STP).
786 The following commands create br0, enable STP, and add eth0
787 and eth1 to the bridge. The order is important because you
788 don't want have to have a loop in your network even
792 ovs-vsctl set bridge br0 stp_enable=true
793 ovs-vsctl add-port br0 eth0
794 ovs-vsctl add-port br0 eth1
796 The Open vSwitch implementation of STP is not well tested.
797 Please report any bugs you observe, but if you'd rather avoid
798 acting as a beta tester then another option might be your
801 ### Q: I can't seem to use Open vSwitch in a wireless network.
803 A: Wireless base stations generally only allow packets with the source
804 MAC address of NIC that completed the initial handshake.
805 Therefore, without MAC rewriting, only a single device can
806 communicate over a single wireless link.
808 This isn't specific to Open vSwitch, it's enforced by the access
809 point, so the same problems will show up with the Linux bridge or
810 any other way to do bridging.
812 ### Q: I can't seem to add my PPP interface to an Open vSwitch bridge.
814 A: PPP most commonly carries IP packets, but Open vSwitch works only
815 with Ethernet frames. The correct way to interface PPP to an
816 Ethernet network is usually to use routing instead of switching.
818 ### Q: Is there any documentation on the database tables and fields?
820 A: Yes. ovs-vswitchd.conf.db(5) is a comprehensive reference.
822 ### Q: When I run ovs-dpctl I no longer see the bridges I created. Instead,
823 I only see a datapath called "ovs-system". How can I see datapath
824 information about a particular bridge?
826 A: In version 1.9.0, OVS switched to using a single datapath that is
827 shared by all bridges of that type. The "ovs-appctl dpif/*"
828 commands provide similar functionality that is scoped by the bridge.
830 ### Q: I created a GRE port using ovs-vsctl so why can't I send traffic or
831 see the port in the datapath?
833 A: On Linux kernels before 3.11, the OVS GRE module and Linux GRE module
834 cannot be loaded at the same time. It is likely that on your system the
835 Linux GRE module is already loaded and blocking OVS (to confirm, check
836 dmesg for errors regarding GRE registration). To fix this, unload all
837 GRE modules that appear in lsmod as well as the OVS kernel module. You
838 can then reload the OVS module following the directions in
839 [INSTALL.md], which will ensure that dependencies are satisfied.
841 ### Q: Open vSwitch does not seem to obey my packet filter rules.
843 A: It depends on mechanisms and configurations you want to use.
845 You cannot usefully use typical packet filters, like iptables, on
846 physical Ethernet ports that you add to an Open vSwitch bridge.
847 This is because Open vSwitch captures packets from the interface at
848 a layer lower below where typical packet-filter implementations
849 install their hooks. (This actually applies to any interface of
850 type "system" that you might add to an Open vSwitch bridge.)
852 You can usefully use typical packet filters on Open vSwitch
853 internal ports as they are mostly ordinary interfaces from the point
854 of view of packet filters.
856 For example, suppose you create a bridge br0 and add Ethernet port
857 eth0 to it. Then you can usefully add iptables rules to affect the
858 internal interface br0, but not the physical interface eth0. (br0
859 is also where you would add an IP address, as discussed elsewhere
862 For simple filtering rules, it might be possible to achieve similar
863 results by installing appropriate OpenFlow flows instead.
865 If the use of a particular packet filter setup is essential, Open
866 vSwitch might not be the best choice for you. On Linux, you might
867 want to consider using the Linux Bridge. (This is the only choice if
868 you want to use ebtables rules.) On NetBSD, you might want to
869 consider using the bridge(4) with BRIDGE_IPF option.
871 ### Q: It seems that Open vSwitch does nothing when I removed a port and
872 then immediately put it back. For example, consider that p1 is
873 a port of type=internal:
875 ovs-vsctl del-port br0 p1 -- \
877 set interface p1 type=internal
879 A: It's an expected behaviour.
881 If del-port and add-port happen in a single OVSDB transaction as
882 your example, Open vSwitch always "skips" the intermediate steps.
883 Even if they are done in multiple transactions, it's still allowed
884 for Open vSwitch to skip the intermediate steps and just implement
885 the overall effect. In both cases, your example would be turned
888 If you want to make Open vSwitch actually destroy and then re-create
889 the port for some side effects like resetting kernel setting for the
890 corresponding interface, you need to separate operations into multiple
891 OVSDB transactions and ensure that at least the first one does not have
892 --no-wait. In the following example, the first ovs-vsctl will block
893 until Open vSwitch reloads the new configuration and removes the port:
895 ovs-vsctl del-port br0 p1
896 ovs-vsctl add-port br0 p1 -- \
897 set interface p1 type=internal
899 ### Q: I want to add thousands of ports to an Open vSwitch bridge, but
900 it takes too long (minutes or hours) to do it with ovs-vsctl. How
903 A: If you add them one at a time with ovs-vsctl, it can take a long
904 time to add thousands of ports to an Open vSwitch bridge. This is
905 because every invocation of ovs-vsctl first reads the current
906 configuration from OVSDB. As the number of ports grows, this
907 starts to take an appreciable amount of time, and when it is
908 repeated thousands of times the total time becomes significant.
910 The solution is to add the ports in one invocation of ovs-vsctl (or
911 a small number of them). For example, using bash:
914 cmds=; for i in {1..5000}; do cmds+=" -- add-port br0 p$i"; done
917 takes seconds, not minutes or hours, in the OVS sandbox environment.
919 ### Q: I created a bridge named br0. My bridge shows up in "ovs-vsctl
920 show", but "ovs-ofctl show br0" just prints "br0 is not a bridge
923 A: Open vSwitch wasn't able to create the bridge. Check the
924 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
925 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log).
927 In general, the Open vSwitch database reflects the desired
928 configuration state. ovs-vswitchd monitors the database and, when
929 it changes, reconfigures the system to reflect the new desired
930 state. This normally happens very quickly. Thus, a discrepancy
931 between the database and the actual state indicates that
932 ovs-vswitchd could not implement the configuration, and so one
933 should check the log to find out why. (Another possible cause is
934 that ovs-vswitchd is not running. This will make "ovs-vsctl"
935 commands hang, if they change the configuration, unless one
936 specifies "--no-wait".)
938 ### Q: I have a bridge br0. I added a new port vif1.0, and it shows
939 up in "ovs-vsctl show", but "ovs-vsctl list port" says that it has
940 OpenFlow port ("ofport") -1, and "ovs-ofctl show br0" doesn't show
943 A: Open vSwitch wasn't able to create the port. Check the
944 ovs-vswitchd log for details (Debian and Red Hat packaging for Open
945 vSwitch put it in /var/log/openvswitch/ovs-vswitchd.log). Please
946 see the previous question for more information.
948 You may want to upgrade to Open vSwitch 2.3 (or later), in which
949 ovs-vsctl will immediately report when there is an issue creating a
952 ### Q: I created a tap device tap0, configured an IP address on it, and
953 added it to a bridge, like this:
956 ifconfig tap0 192.168.0.123
958 ovs-vsctl add-port br0 tap0
960 I expected that I could then use this IP address to contact other
961 hosts on the network, but it doesn't work. Why not?
963 A: The short answer is that this is a misuse of a "tap" device. Use
964 an "internal" device implemented by Open vSwitch, which works
965 differently and is designed for this use. To solve this problem
966 with an internal device, instead run:
969 ovs-vsctl add-port br0 int0 -- set Interface int0 type=internal
970 ifconfig int0 192.168.0.123
972 Even more simply, you can take advantage of the internal port that
973 every bridge has under the name of the bridge:
976 ifconfig br0 192.168.0.123
978 In more detail, a "tap" device is an interface between the Linux
979 (or *BSD) network stack and a user program that opens it as a
980 socket. When the "tap" device transmits a packet, it appears in
981 the socket opened by the userspace program. Conversely, when the
982 userspace program writes to the "tap" socket, the kernel TCP/IP
983 stack processes the packet as if it had been received by the "tap"
986 Consider the configuration above. Given this configuration, if you
987 "ping" an IP address in the 192.168.0.x subnet, the Linux kernel
988 routing stack will transmit an ARP on the tap0 device. Open
989 vSwitch userspace treats "tap" devices just like any other network
990 device; that is, it doesn't open them as "tap" sockets. That means
991 that the ARP packet will simply get dropped.
993 You might wonder why the Open vSwitch kernel module doesn't
994 intercept the ARP packet and bridge it. After all, Open vSwitch
995 intercepts packets on other devices. The answer is that Open
996 vSwitch only intercepts *received* packets, but this is a packet
997 being transmitted. The same thing happens for all other types of
998 network devices, except for Open vSwitch "internal" ports. If you,
999 for example, add a physical Ethernet port to an OVS bridge,
1000 configure an IP address on a physical Ethernet port, and then issue
1001 a "ping" to an address in that subnet, the same thing happens: an
1002 ARP gets transmitted on the physical Ethernet port and Open vSwitch
1003 never sees it. (You should not do that, as documented at the
1004 beginning of this section.)
1006 It can make sense to add a "tap" device to an Open vSwitch bridge,
1007 if some userspace program (other than Open vSwitch) has opened the
1008 tap socket. This is the case, for example, if the "tap" device was
1009 created by KVM (or QEMU) to simulate a virtual NIC. In such a
1010 case, when OVS bridges a packet to the "tap" device, the kernel
1011 forwards that packet to KVM in userspace, which passes it along to
1012 the VM, and in the other direction, when the VM sends a packet, KVM
1013 writes it to the "tap" socket, which causes OVS to receive it and
1014 bridge it to the other OVS ports. Please note that in such a case
1015 no IP address is configured on the "tap" device (there is normally
1016 an IP address configured in the virtual NIC inside the VM, but this
1017 is not visible to the host Linux kernel or to Open vSwitch).
1019 There is one special case in which Open vSwitch does directly read
1020 and write "tap" sockets. This is an implementation detail of the
1021 Open vSwitch userspace switch, which implements its "internal"
1022 ports as Linux (or *BSD) "tap" sockets. In such a userspace
1023 switch, OVS receives packets sent on the "tap" device used to
1024 implement an "internal" port by reading the associated "tap"
1025 socket, and bridges them to the rest of the switch. In the other
1026 direction, OVS transmits packets bridged to the "internal" port by
1027 writing them to the "tap" socket, causing them to be processed by
1028 the kernel TCP/IP stack as if they had been received on the "tap"
1029 device. Users should not need to be concerned with this
1030 implementation detail.
1032 Open vSwitch has a network device type called "tap". This is
1033 intended only for implementing "internal" ports in the OVS
1034 userspace switch and should not be used otherwise. In particular,
1035 users should not configure KVM "tap" devices as type "tap" (use
1036 type "system", the default, instead).
1039 Quality of Service (QoS)
1040 ------------------------
1042 ### Q: Does OVS support Quality of Service (QoS)?
1044 A: Yes. For traffic that egresses from a switch, OVS supports traffic
1045 shaping; for traffic that ingresses into a switch, OVS support
1046 policing. Policing is a simple form of quality-of-service that
1047 simply drops packets received in excess of the configured rate. Due
1048 to its simplicity, policing is usually less accurate and less
1049 effective than egress traffic shaping, which queues packets.
1051 Keep in mind that ingress and egress are from the perspective of the
1052 switch. That means that egress shaping limits the rate at which
1053 traffic is allowed to transmit from a physical interface, but the
1054 rate at which traffic will be received on a virtual machine's VIF.
1055 For ingress policing, the behavior is the opposite.
1057 ### Q: How do I configure egress traffic shaping?
1059 A: Suppose that you want to set up bridge br0 connected to physical
1060 Ethernet port eth0 (a 1 Gbps device) and virtual machine interfaces
1061 vif1.0 and vif2.0, and that you want to limit traffic from vif1.0
1062 to eth0 to 10 Mbps and from vif2.0 to eth0 to 20 Mbps. Then, you
1063 could configure the bridge this way:
1067 add-port br0 eth0 -- \
1068 add-port br0 vif1.0 -- set interface vif1.0 ofport_request=5 -- \
1069 add-port br0 vif2.0 -- set interface vif2.0 ofport_request=6 -- \
1070 set port eth0 qos=@newqos -- \
1071 --id=@newqos create qos type=linux-htb \
1072 other-config:max-rate=1000000000 \
1073 queues:123=@vif10queue \
1074 queues:234=@vif20queue -- \
1075 --id=@vif10queue create queue other-config:max-rate=10000000 -- \
1076 --id=@vif20queue create queue other-config:max-rate=20000000
1078 At this point, bridge br0 is configured with the ports and eth0 is
1079 configured with the queues that you need for QoS, but nothing is
1080 actually directing packets from vif1.0 or vif2.0 to the queues that
1081 we have set up for them. That means that all of the packets to
1082 eth0 are going to the "default queue", which is not what we want.
1084 We use OpenFlow to direct packets from vif1.0 and vif2.0 to the
1085 queues reserved for them:
1087 ovs-ofctl add-flow br0 in_port=5,actions=set_queue:123,normal
1088 ovs-ofctl add-flow br0 in_port=6,actions=set_queue:234,normal
1090 Each of the above flows matches on the input port, sets up the
1091 appropriate queue (123 for vif1.0, 234 for vif2.0), and then
1092 executes the "normal" action, which performs the same switching
1093 that Open vSwitch would have done without any OpenFlow flows being
1094 present. (We know that vif1.0 and vif2.0 have OpenFlow port
1095 numbers 5 and 6, respectively, because we set their ofport_request
1096 columns above. If we had not done that, then we would have needed
1097 to find out their port numbers before setting up these flows.)
1099 Now traffic going from vif1.0 or vif2.0 to eth0 should be
1102 By the way, if you delete the bridge created by the above commands,
1105 ovs-vsctl del-br br0
1107 then that will leave one unreferenced QoS record and two
1108 unreferenced Queue records in the Open vSwich database. One way to
1109 clear them out, assuming you don't have other QoS or Queue records
1110 that you want to keep, is:
1112 ovs-vsctl -- --all destroy QoS -- --all destroy Queue
1114 If you do want to keep some QoS or Queue records, or the Open
1115 vSwitch you are using is older than version 1.8 (which added the
1116 --all option), then you will have to destroy QoS and Queue records
1119 ### Q: How do I configure ingress policing?
1121 A: A policing policy can be configured on an interface to drop packets
1122 that arrive at a higher rate than the configured value. For example,
1123 the following commands will rate-limit traffic that vif1.0 may
1126 ovs-vsctl set interface vif1.0 ingress_policing_rate=10000
1127 ovs-vsctl set interface vif1.0 ingress_policing_burst=1000
1129 Traffic policing can interact poorly with some network protocols and
1130 can have surprising results. The "Ingress Policing" section of
1131 ovs-vswitchd.conf.db(5) discusses the issues in greater detail.
1133 ### Q: I configured Quality of Service (QoS) in my OpenFlow network by
1134 adding records to the QoS and Queue table, but the results aren't
1137 A: Did you install OpenFlow flows that use your queues? This is the
1138 primary way to tell Open vSwitch which queues you want to use. If
1139 you don't do this, then the default queue will be used, which will
1140 probably not have the effect you want.
1142 Refer to the previous question for an example.
1144 ### Q: I'd like to take advantage of some QoS feature that Open vSwitch
1145 doesn't yet support. How do I do that?
1147 A: Open vSwitch does not implement QoS itself. Instead, it can
1148 configure some, but not all, of the QoS features built into the
1149 Linux kernel. If you need some QoS feature that OVS cannot
1150 configure itself, then the first step is to figure out whether
1151 Linux QoS supports that feature. If it does, then you can submit a
1152 patch to support Open vSwitch configuration for that feature, or
1153 you can use "tc" directly to configure the feature in Linux. (If
1154 Linux QoS doesn't support the feature you want, then first you have
1155 to add that support to Linux.)
1157 ### Q: I configured QoS, correctly, but my measurements show that it isn't
1158 working as well as I expect.
1160 A: With the Linux kernel, the Open vSwitch implementation of QoS has
1163 - Open vSwitch configures a subset of Linux kernel QoS
1164 features, according to what is in OVSDB. It is possible that
1165 this code has bugs. If you believe that this is so, then you
1166 can configure the Linux traffic control (QoS) stack directly
1167 with the "tc" program. If you get better results that way,
1168 you can send a detailed bug report to bugs@openvswitch.org.
1170 It is certain that Open vSwitch cannot configure every Linux
1171 kernel QoS feature. If you need some feature that OVS cannot
1172 configure, then you can also use "tc" directly (or add that
1175 - The Open vSwitch implementation of OpenFlow allows flows to
1176 be directed to particular queues. This is pretty simple and
1177 unlikely to have serious bugs at this point.
1179 However, most problems with QoS on Linux are not bugs in Open
1180 vSwitch at all. They tend to be either configuration errors
1181 (please see the earlier questions in this section) or issues with
1182 the traffic control (QoS) stack in Linux. The Open vSwitch
1183 developers are not experts on Linux traffic control. We suggest
1184 that, if you believe you are encountering a problem with Linux
1185 traffic control, that you consult the tc manpages (e.g. tc(8),
1186 tc-htb(8), tc-hfsc(8)), web resources (e.g. http://lartc.org/), or
1187 mailing lists (e.g. http://vger.kernel.org/vger-lists.html#netdev).
1189 ### Q: Does Open vSwitch support OpenFlow meters?
1191 A: Since version 2.0, Open vSwitch has OpenFlow protocol support for
1192 OpenFlow meters. There is no implementation of meters in the Open
1193 vSwitch software switch (neither the kernel-based nor userspace
1200 ### Q: What's a VLAN?
1202 A: At the simplest level, a VLAN (short for "virtual LAN") is a way to
1203 partition a single switch into multiple switches. Suppose, for
1204 example, that you have two groups of machines, group A and group B.
1205 You want the machines in group A to be able to talk to each other,
1206 and you want the machine in group B to be able to talk to each
1207 other, but you don't want the machines in group A to be able to
1208 talk to the machines in group B. You can do this with two
1209 switches, by plugging the machines in group A into one switch and
1210 the machines in group B into the other switch.
1212 If you only have one switch, then you can use VLANs to do the same
1213 thing, by configuring the ports for machines in group A as VLAN
1214 "access ports" for one VLAN and the ports for group B as "access
1215 ports" for a different VLAN. The switch will only forward packets
1216 between ports that are assigned to the same VLAN, so this
1217 effectively subdivides your single switch into two independent
1218 switches, one for each group of machines.
1220 So far we haven't said anything about VLAN headers. With access
1221 ports, like we've described so far, no VLAN header is present in
1222 the Ethernet frame. This means that the machines (or switches)
1223 connected to access ports need not be aware that VLANs are
1224 involved, just like in the case where we use two different physical
1227 Now suppose that you have a whole bunch of switches in your
1228 network, instead of just one, and that some machines in group A are
1229 connected directly to both switches 1 and 2. To allow these
1230 machines to talk to each other, you could add an access port for
1231 group A's VLAN to switch 1 and another to switch 2, and then
1232 connect an Ethernet cable between those ports. That works fine,
1233 but it doesn't scale well as the number of switches and the number
1234 of VLANs increases, because you use up a lot of valuable switch
1235 ports just connecting together your VLANs.
1237 This is where VLAN headers come in. Instead of using one cable and
1238 two ports per VLAN to connect a pair of switches, we configure a
1239 port on each switch as a VLAN "trunk port". Packets sent and
1240 received on a trunk port carry a VLAN header that says what VLAN
1241 the packet belongs to, so that only two ports total are required to
1242 connect the switches, regardless of the number of VLANs in use.
1243 Normally, only switches (either physical or virtual) are connected
1244 to a trunk port, not individual hosts, because individual hosts
1245 don't expect to see a VLAN header in the traffic that they receive.
1247 None of the above discussion says anything about particular VLAN
1248 numbers. This is because VLAN numbers are completely arbitrary.
1249 One must only ensure that a given VLAN is numbered consistently
1250 throughout a network and that different VLANs are given different
1251 numbers. (That said, VLAN 0 is usually synonymous with a packet
1252 that has no VLAN header, and VLAN 4095 is reserved.)
1254 ### Q: VLANs don't work.
1256 A: Many drivers in Linux kernels before version 3.3 had VLAN-related
1257 bugs. If you are having problems with VLANs that you suspect to be
1258 driver related, then you have several options:
1260 - Upgrade to Linux 3.3 or later.
1262 - Build and install a fixed version of the particular driver
1263 that is causing trouble, if one is available.
1265 - Use a NIC whose driver does not have VLAN problems.
1267 - Use "VLAN splinters", a feature in Open vSwitch 1.4 and later
1268 that works around bugs in kernel drivers. To enable VLAN
1269 splinters on interface eth0, use the command:
1271 ovs-vsctl set interface eth0 other-config:enable-vlan-splinters=true
1273 For VLAN splinters to be effective, Open vSwitch must know
1274 which VLANs are in use. See the "VLAN splinters" section in
1275 the Interface table in ovs-vswitchd.conf.db(5) for details on
1276 how Open vSwitch infers in-use VLANs.
1278 VLAN splinters increase memory use and reduce performance, so
1279 use them only if needed.
1281 - Apply the "vlan workaround" patch from the XenServer kernel
1282 patch queue, build Open vSwitch against this patched kernel,
1283 and then use ovs-vlan-bug-workaround(8) to enable the VLAN
1284 workaround for each interface whose driver is buggy.
1286 (This is a nontrivial exercise, so this option is included
1287 only for completeness.)
1289 It is not always easy to tell whether a Linux kernel driver has
1290 buggy VLAN support. The ovs-vlan-test(8) and ovs-test(8) utilities
1291 can help you test. See their manpages for details. Of the two
1292 utilities, ovs-test(8) is newer and more thorough, but
1293 ovs-vlan-test(8) may be easier to use.
1295 ### Q: VLANs still don't work. I've tested the driver so I know that it's OK.
1297 A: Do you have VLANs enabled on the physical switch that OVS is
1298 attached to? Make sure that the port is configured to trunk the
1299 VLAN or VLANs that you are using with OVS.
1301 ### Q: Outgoing VLAN-tagged traffic goes through OVS to my physical switch
1302 and to its destination host, but OVS seems to drop incoming return
1305 A: It's possible that you have the VLAN configured on your physical
1306 switch as the "native" VLAN. In this mode, the switch treats
1307 incoming packets either tagged with the native VLAN or untagged as
1308 part of the native VLAN. It may also send outgoing packets in the
1309 native VLAN without a VLAN tag.
1311 If this is the case, you have two choices:
1313 - Change the physical switch port configuration to tag packets
1314 it forwards to OVS with the native VLAN instead of forwarding
1317 - Change the OVS configuration for the physical port to a
1318 native VLAN mode. For example, the following sets up a
1319 bridge with port eth0 in "native-tagged" mode in VLAN 9:
1321 ovs-vsctl add-br br0
1322 ovs-vsctl add-port br0 eth0 tag=9 vlan_mode=native-tagged
1324 In this situation, "native-untagged" mode will probably work
1325 equally well. Refer to the documentation for the Port table
1326 in ovs-vswitchd.conf.db(5) for more information.
1328 ### Q: I added a pair of VMs on different VLANs, like this:
1330 ovs-vsctl add-br br0
1331 ovs-vsctl add-port br0 eth0
1332 ovs-vsctl add-port br0 tap0 tag=9
1333 ovs-vsctl add-port br0 tap1 tag=10
1335 but the VMs can't access each other, the external network, or the
1338 A: It is to be expected that the VMs can't access each other. VLANs
1339 are a means to partition a network. When you configured tap0 and
1340 tap1 as access ports for different VLANs, you indicated that they
1341 should be isolated from each other.
1343 As for the external network and the Internet, it seems likely that
1344 the machines you are trying to access are not on VLAN 9 (or 10) and
1345 that the Internet is not available on VLAN 9 (or 10).
1347 ### Q: I added a pair of VMs on the same VLAN, like this:
1349 ovs-vsctl add-br br0
1350 ovs-vsctl add-port br0 eth0
1351 ovs-vsctl add-port br0 tap0 tag=9
1352 ovs-vsctl add-port br0 tap1 tag=9
1354 The VMs can access each other, but not the external network or the
1357 A: It seems likely that the machines you are trying to access in the
1358 external network are not on VLAN 9 and that the Internet is not
1359 available on VLAN 9. Also, ensure VLAN 9 is set up as an allowed
1360 trunk VLAN on the upstream switch port to which eth0 is connected.
1362 ### Q: Can I configure an IP address on a VLAN?
1364 A: Yes. Use an "internal port" configured as an access port. For
1365 example, the following configures IP address 192.168.0.7 on VLAN 9.
1366 That is, OVS will forward packets from eth0 to 192.168.0.7 only if
1367 they have an 802.1Q header with VLAN 9. Conversely, traffic
1368 forwarded from 192.168.0.7 to eth0 will be tagged with an 802.1Q
1371 ovs-vsctl add-br br0
1372 ovs-vsctl add-port br0 eth0
1373 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1374 ifconfig vlan9 192.168.0.7
1376 See also the following question.
1378 ### Q: I configured one IP address on VLAN 0 and another on VLAN 9, like
1381 ovs-vsctl add-br br0
1382 ovs-vsctl add-port br0 eth0
1383 ifconfig br0 192.168.0.5
1384 ovs-vsctl add-port br0 vlan9 tag=9 -- set interface vlan9 type=internal
1385 ifconfig vlan9 192.168.0.9
1387 but other hosts that are only on VLAN 0 can reach the IP address
1388 configured on VLAN 9. What's going on?
1390 A: RFC 1122 section 3.3.4.2 "Multihoming Requirements" describes two
1391 approaches to IP address handling in Internet hosts:
1393 - In the "Strong ES Model", where an ES is a host ("End
1394 System"), an IP address is primarily associated with a
1395 particular interface. The host discards packets that arrive
1396 on interface A if they are destined for an IP address that is
1397 configured on interface B. The host never sends packets from
1398 interface A using a source address configured on interface B.
1400 - In the "Weak ES Model", an IP address is primarily associated
1401 with a host. The host accepts packets that arrive on any
1402 interface if they are destined for any of the host's IP
1403 addresses, even if the address is configured on some
1404 interface other than the one on which it arrived. The host
1405 does not restrict itself to sending packets from an IP
1406 address associated with the originating interface.
1408 Linux uses the weak ES model. That means that when packets
1409 destined to the VLAN 9 IP address arrive on eth0 and are bridged to
1410 br0, the kernel IP stack accepts them there for the VLAN 9 IP
1411 address, even though they were not received on vlan9, the network
1414 To simulate the strong ES model on Linux, one may add iptables rule
1415 to filter packets based on source and destination address and
1416 adjust ARP configuration with sysctls.
1418 BSD uses the strong ES model.
1420 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1422 A: The configuration for VLANs in the Open vSwitch database (e.g. via
1423 ovs-vsctl) only affects traffic that goes through Open vSwitch's
1424 implementation of the OpenFlow "normal switching" action. By
1425 default, when Open vSwitch isn't connected to a controller and
1426 nothing has been manually configured in the flow table, all traffic
1427 goes through the "normal switching" action. But, if you set up
1428 OpenFlow flows on your own, through a controller or using ovs-ofctl
1429 or through other means, then you have to implement VLAN handling
1432 You can use "normal switching" as a component of your OpenFlow
1433 actions, e.g. by putting "normal" into the lists of actions on
1434 ovs-ofctl or by outputting to OFPP_NORMAL from an OpenFlow
1435 controller. In situations where this is not suitable, you can
1436 implement VLAN handling yourself, e.g.:
1438 - If a packet comes in on an access port, and the flow table
1439 needs to send it out on a trunk port, then the flow can add
1440 the appropriate VLAN tag with the "mod_vlan_vid" action.
1442 - If a packet comes in on a trunk port, and the flow table
1443 needs to send it out on an access port, then the flow can
1444 strip the VLAN tag with the "strip_vlan" action.
1446 ### Q: I configured ports on a bridge as access ports with different VLAN
1449 ovs-vsctl add-br br0
1450 ovs-vsctl set-controller br0 tcp:192.168.0.10:6653
1451 ovs-vsctl add-port br0 eth0
1452 ovs-vsctl add-port br0 tap0 tag=9
1453 ovs-vsctl add-port br0 tap1 tag=10
1455 but the VMs running behind tap0 and tap1 can still communicate,
1456 that is, they are not isolated from each other even though they are
1459 A: Do you have a controller configured on br0 (as the commands above
1460 do)? If so, then this is a variant on the previous question, "My
1461 OpenFlow controller doesn't see the VLANs that I expect," and you
1462 can refer to the answer there for more information.
1464 ### Q: How MAC learning works with VLANs?
1466 A: Open vSwitch implements Independent VLAN Learning (IVL) for
1467 OFPP_NORMAL action. I.e. it logically has separate learning tables
1474 ### Q: What's a VXLAN?
1476 A: VXLAN stands for Virtual eXtensible Local Area Network, and is a means
1477 to solve the scaling challenges of VLAN networks in a multi-tenant
1478 environment. VXLAN is an overlay network which transports an L2 network
1479 over an existing L3 network. For more information on VXLAN, please see
1482 http://tools.ietf.org/html/rfc7348
1484 ### Q: How much of the VXLAN protocol does Open vSwitch currently support?
1486 A: Open vSwitch currently supports the framing format for packets on the
1487 wire. There is currently no support for the multicast aspects of VXLAN.
1488 To get around the lack of multicast support, it is possible to
1489 pre-provision MAC to IP address mappings either manually or from a
1492 ### Q: What destination UDP port does the VXLAN implementation in Open vSwitch
1495 A: By default, Open vSwitch will use the assigned IANA port for VXLAN, which
1496 is 4789. However, it is possible to configure the destination UDP port
1497 manually on a per-VXLAN tunnel basis. An example of this configuration is
1500 ovs-vsctl add-br br0
1501 ovs-vsctl add-port br0 vxlan1 -- set interface vxlan1
1502 type=vxlan options:remote_ip=192.168.1.2 options:key=flow
1503 options:dst_port=8472
1506 Using OpenFlow (Manually or Via Controller)
1507 -------------------------------------------
1509 ### Q: What versions of OpenFlow does Open vSwitch support?
1511 A: The following table lists the versions of OpenFlow supported by
1512 each version of Open vSwitch:
1514 Open vSwitch OF1.0 OF1.1 OF1.2 OF1.3 OF1.4 OF1.5
1515 ###============ ===== ===== ===== ===== ===== =====
1516 1.9 and earlier yes --- --- --- --- ---
1517 1.10 yes --- [*] [*] --- ---
1518 1.11 yes --- [*] [*] --- ---
1519 2.0 yes [*] [*] [*] --- ---
1520 2.1 yes [*] [*] [*] --- ---
1521 2.2 yes [*] [*] [*] [%] [*]
1522 2.3 yes yes yes yes [*] [*]
1524 [*] Supported, with one or more missing features.
1525 [%] Experimental, unsafe implementation.
1527 Open vSwitch 2.3 enables OpenFlow 1.0, 1.1, 1.2, and 1.3 by default
1528 in ovs-vswitchd. In Open vSwitch 1.10 through 2.2, OpenFlow 1.1,
1529 1.2, and 1.3 must be enabled manually in ovs-vswitchd. OpenFlow
1530 1.4 and 1.5 are also supported, with missing features, in Open
1531 vSwitch 2.3 and later, but not enabled by default. In any case,
1532 the user may override the default:
1534 - To enable OpenFlow 1.0, 1.1, 1.2, and 1.3 on bridge br0:
1536 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13
1538 - To enable OpenFlow 1.0, 1.1, 1.2, 1.3, 1.4, and 1.5 on bridge br0:
1540 ovs-vsctl set bridge br0 protocols=OpenFlow10,OpenFlow11,OpenFlow12,OpenFlow13,OpenFlow14,OpenFlow15
1542 - To enable only OpenFlow 1.0 on bridge br0:
1544 ovs-vsctl set bridge br0 protocols=OpenFlow10
1546 All current versions of ovs-ofctl enable only OpenFlow 1.0 by
1547 default. Use the -O option to enable support for later versions of
1548 OpenFlow in ovs-ofctl. For example:
1550 ovs-ofctl -O OpenFlow13 dump-flows br0
1552 (Open vSwitch 2.2 had an experimental implementation of OpenFlow
1553 1.4 that could cause crashes. We don't recommend enabling it.)
1555 [OPENFLOW-1.1+.md] in the Open vSwitch source tree tracks support for
1556 OpenFlow 1.1 and later features. When support for OpenFlow 1.4 and
1557 1.5 is solidly implemented, Open vSwitch will enable those version
1560 ### Q: Does Open vSwitch support MPLS?
1562 A: Before version 1.11, Open vSwitch did not support MPLS. That is,
1563 these versions can match on MPLS Ethernet types, but they cannot
1564 match, push, or pop MPLS labels, nor can they look past MPLS labels
1565 into the encapsulated packet.
1567 Open vSwitch versions 1.11, 2.0, and 2.1 have very minimal support
1568 for MPLS. With the userspace datapath only, these versions can
1569 match, push, or pop a single MPLS label, but they still cannot look
1570 past MPLS labels (even after popping them) into the encapsulated
1571 packet. Kernel datapath support is unchanged from earlier
1574 Open vSwitch version 2.3 can match, push, or pop a single MPLS
1575 label and look past the MPLS label into the encapsulated packet.
1576 Both userspace and kernel datapaths will be supported, but MPLS
1577 processing always happens in userspace either way, so kernel
1578 datapath performance will be disappointing.
1580 Open vSwitch version 2.4 can match, push, or pop up to 3 MPLS
1581 labels and look past the MPLS label into the encapsulated packet.
1582 It will have kernel support for MPLS, yielding improved
1585 ### Q: I'm getting "error type 45250 code 0". What's that?
1587 A: This is a Open vSwitch extension to OpenFlow error codes. Open
1588 vSwitch uses this extension when it must report an error to an
1589 OpenFlow controller but no standard OpenFlow error code is
1592 Open vSwitch logs the errors that it sends to controllers, so the
1593 easiest thing to do is probably to look at the ovs-vswitchd log to
1594 find out what the error was.
1596 If you want to dissect the extended error message yourself, the
1597 format is documented in include/openflow/nicira-ext.h in the Open
1598 vSwitch source distribution. The extended error codes are
1599 documented in include/openvswitch/ofp-errors.h.
1601 Q1: Some of the traffic that I'd expect my OpenFlow controller to see
1602 doesn't actually appear through the OpenFlow connection, even
1603 though I know that it's going through.
1604 Q2: Some of the OpenFlow flows that my controller sets up don't seem
1605 to apply to certain traffic, especially traffic between OVS and
1606 the controller itself.
1608 A: By default, Open vSwitch assumes that OpenFlow controllers are
1609 connected "in-band", that is, that the controllers are actually
1610 part of the network that is being controlled. In in-band mode,
1611 Open vSwitch sets up special "hidden" flows to make sure that
1612 traffic can make it back and forth between OVS and the controllers.
1613 These hidden flows are higher priority than any flows that can be
1614 set up through OpenFlow, and they are not visible through normal
1615 OpenFlow flow table dumps.
1617 Usually, the hidden flows are desirable and helpful, but
1618 occasionally they can cause unexpected behavior. You can view the
1619 full OpenFlow flow table, including hidden flows, on bridge br0
1622 ovs-appctl bridge/dump-flows br0
1624 to help you debug. The hidden flows are those with priorities
1625 greater than 65535 (the maximum priority that can be set with
1628 The DESIGN file at the top level of the Open vSwitch source
1629 distribution describes the in-band model in detail.
1631 If your controllers are not actually in-band (e.g. they are on
1632 localhost via 127.0.0.1, or on a separate network), then you should
1633 configure your controllers in "out-of-band" mode. If you have one
1634 controller on bridge br0, then you can configure out-of-band mode
1637 ovs-vsctl set controller br0 connection-mode=out-of-band
1639 ### Q: I configured all my controllers for out-of-band control mode but
1640 "ovs-appctl bridge/dump-flows" still shows some hidden flows.
1642 A: You probably have a remote manager configured (e.g. with "ovs-vsctl
1643 set-manager"). By default, Open vSwitch assumes that managers need
1644 in-band rules set up on every bridge. You can disable these rules
1647 ovs-vsctl set bridge br0 other-config:disable-in-band=true
1649 This actually disables in-band control entirely for the bridge, as
1650 if all the bridge's controllers were configured for out-of-band
1653 ### Q: My OpenFlow controller doesn't see the VLANs that I expect.
1655 A: See answer under "VLANs", above.
1657 ### Q: I ran "ovs-ofctl add-flow br0 nw_dst=192.168.0.1,actions=drop"
1658 but I got a funny message like this:
1660 ofp_util|INFO|normalization changed ofp_match, details:
1661 ofp_util|INFO| pre: nw_dst=192.168.0.1
1664 and when I ran "ovs-ofctl dump-flows br0" I saw that my nw_dst
1665 match had disappeared, so that the flow ends up matching every
1668 A: The term "normalization" in the log message means that a flow
1669 cannot match on an L3 field without saying what L3 protocol is in
1670 use. The "ovs-ofctl" command above didn't specify an L3 protocol,
1671 so the L3 field match was dropped.
1673 In this case, the L3 protocol could be IP or ARP. A correct
1674 command for each possibility is, respectively:
1676 ovs-ofctl add-flow br0 ip,nw_dst=192.168.0.1,actions=drop
1680 ovs-ofctl add-flow br0 arp,nw_dst=192.168.0.1,actions=drop
1682 Similarly, a flow cannot match on an L4 field without saying what
1683 L4 protocol is in use. For example, the flow match "tp_src=1234"
1684 is, by itself, meaningless and will be ignored. Instead, to match
1685 TCP source port 1234, write "tcp,tp_src=1234", or to match UDP
1686 source port 1234, write "udp,tp_src=1234".
1688 ### Q: How can I figure out the OpenFlow port number for a given port?
1690 A: The OFPT_FEATURES_REQUEST message requests an OpenFlow switch to
1691 respond with an OFPT_FEATURES_REPLY that, among other information,
1692 includes a mapping between OpenFlow port names and numbers. From a
1693 command prompt, "ovs-ofctl show br0" makes such a request and
1694 prints the response for switch br0.
1696 The Interface table in the Open vSwitch database also maps OpenFlow
1697 port names to numbers. To print the OpenFlow port number
1698 associated with interface eth0, run:
1700 ovs-vsctl get Interface eth0 ofport
1702 You can print the entire mapping with:
1704 ovs-vsctl -- --columns=name,ofport list Interface
1706 but the output mixes together interfaces from all bridges in the
1707 database, so it may be confusing if more than one bridge exists.
1709 In the Open vSwitch database, ofport value -1 means that the
1710 interface could not be created due to an error. (The Open vSwitch
1711 log should indicate the reason.) ofport value [] (the empty set)
1712 means that the interface hasn't been created yet. The latter is
1713 normally an intermittent condition (unless ovs-vswitchd is not
1716 ### Q: I added some flows with my controller or with ovs-ofctl, but when I
1717 run "ovs-dpctl dump-flows" I don't see them.
1719 A: ovs-dpctl queries a kernel datapath, not an OpenFlow switch. It
1720 won't display the information that you want. You want to use
1721 "ovs-ofctl dump-flows" instead.
1723 ### Q: It looks like each of the interfaces in my bonded port shows up
1724 as an individual OpenFlow port. Is that right?
1726 A: Yes, Open vSwitch makes individual bond interfaces visible as
1727 OpenFlow ports, rather than the bond as a whole. The interfaces
1728 are treated together as a bond for only a few purposes:
1730 - Sending a packet to the OFPP_NORMAL port. (When an OpenFlow
1731 controller is not configured, this happens implicitly to
1734 - Mirrors configured for output to a bonded port.
1736 It would make a lot of sense for Open vSwitch to present a bond as
1737 a single OpenFlow port. If you want to contribute an
1738 implementation of such a feature, please bring it up on the Open
1739 vSwitch development mailing list at dev@openvswitch.org.
1741 ### Q: I have a sophisticated network setup involving Open vSwitch, VMs or
1742 multiple hosts, and other components. The behavior isn't what I
1745 A: To debug network behavior problems, trace the path of a packet,
1746 hop-by-hop, from its origin in one host to a remote host. If
1747 that's correct, then trace the path of the response packet back to
1750 The open source tool called "plotnetcfg" can help to understand the
1751 relationship between the networking devices on a single host.
1753 Usually a simple ICMP echo request and reply ("ping") packet is
1754 good enough. Start by initiating an ongoing "ping" from the origin
1755 host to a remote host. If you are tracking down a connectivity
1756 problem, the "ping" will not display any successful output, but
1757 packets are still being sent. (In this case the packets being sent
1758 are likely ARP rather than ICMP.)
1760 Tools available for tracing include the following:
1762 - "tcpdump" and "wireshark" for observing hops across network
1763 devices, such as Open vSwitch internal devices and physical
1766 - "ovs-appctl dpif/dump-flows <br>" in Open vSwitch 1.10 and
1767 later or "ovs-dpctl dump-flows <br>" in earlier versions.
1768 These tools allow one to observe the actions being taken on
1769 packets in ongoing flows.
1771 See ovs-vswitchd(8) for "ovs-appctl dpif/dump-flows"
1772 documentation, ovs-dpctl(8) for "ovs-dpctl dump-flows"
1773 documentation, and "Why are there so many different ways to
1774 dump flows?" above for some background.
1776 - "ovs-appctl ofproto/trace" to observe the logic behind how
1777 ovs-vswitchd treats packets. See ovs-vswitchd(8) for
1778 documentation. You can out more details about a given flow
1779 that "ovs-dpctl dump-flows" displays, by cutting and pasting
1780 a flow from the output into an "ovs-appctl ofproto/trace"
1783 - SPAN, RSPAN, and ERSPAN features of physical switches, to
1784 observe what goes on at these physical hops.
1786 Starting at the origin of a given packet, observe the packet at
1787 each hop in turn. For example, in one plausible scenario, you
1790 1. "tcpdump" the "eth" interface through which an ARP egresses
1791 a VM, from inside the VM.
1793 2. "tcpdump" the "vif" or "tap" interface through which the ARP
1794 ingresses the host machine.
1796 3. Use "ovs-dpctl dump-flows" to spot the ARP flow and observe
1797 the host interface through which the ARP egresses the
1798 physical machine. You may need to use "ovs-dpctl show" to
1799 interpret the port numbers. If the output seems surprising,
1800 you can use "ovs-appctl ofproto/trace" to observe details of
1801 how ovs-vswitchd determined the actions in the "ovs-dpctl
1804 4. "tcpdump" the "eth" interface through which the ARP egresses
1805 the physical machine.
1807 5. "tcpdump" the "eth" interface through which the ARP
1808 ingresses the physical machine, at the remote host that
1811 6. Use "ovs-dpctl dump-flows" to spot the ARP flow on the
1812 remote host that receives the ARP and observe the VM "vif"
1813 or "tap" interface to which the flow is directed. Again,
1814 "ovs-dpctl show" and "ovs-appctl ofproto/trace" might help.
1816 7. "tcpdump" the "vif" or "tap" interface to which the ARP is
1819 8. "tcpdump" the "eth" interface through which the ARP
1820 ingresses a VM, from inside the VM.
1822 It is likely that during one of these steps you will figure out the
1823 problem. If not, then follow the ARP reply back to the origin, in
1826 ### Q: How do I make a flow drop packets?
1828 A: To drop a packet is to receive it without forwarding it. OpenFlow
1829 explicitly specifies forwarding actions. Thus, a flow with an
1830 empty set of actions does not forward packets anywhere, causing
1831 them to be dropped. You can specify an empty set of actions with
1832 "actions=" on the ovs-ofctl command line. For example:
1834 ovs-ofctl add-flow br0 priority=65535,actions=
1836 would cause every packet entering switch br0 to be dropped.
1838 You can write "drop" explicitly if you like. The effect is the
1839 same. Thus, the following command also causes every packet
1840 entering switch br0 to be dropped:
1842 ovs-ofctl add-flow br0 priority=65535,actions=drop
1844 "drop" is not an action, either in OpenFlow or Open vSwitch.
1845 Rather, it is only a way to say that there are no actions.
1847 ### Q: I added a flow to send packets out the ingress port, like this:
1849 ovs-ofctl add-flow br0 in_port=2,actions=2
1851 but OVS drops the packets instead.
1853 A: Yes, OpenFlow requires a switch to ignore attempts to send a packet
1854 out its ingress port. The rationale is that dropping these packets
1855 makes it harder to loop the network. Sometimes this behavior can
1856 even be convenient, e.g. it is often the desired behavior in a flow
1857 that forwards a packet to several ports ("floods" the packet).
1859 Sometimes one really needs to send a packet out its ingress port
1860 ("hairpin"). In this case, output to OFPP_IN_PORT, which in
1861 ovs-ofctl syntax is expressed as just "in_port", e.g.:
1863 ovs-ofctl add-flow br0 in_port=2,actions=in_port
1865 This also works in some circumstances where the flow doesn't match
1866 on the input port. For example, if you know that your switch has
1867 five ports numbered 2 through 6, then the following will send every
1868 received packet out every port, even its ingress port:
1870 ovs-ofctl add-flow br0 actions=2,3,4,5,6,in_port
1874 ovs-ofctl add-flow br0 actions=all,in_port
1876 Sometimes, in complicated flow tables with multiple levels of
1877 "resubmit" actions, a flow needs to output to a particular port
1878 that may or may not be the ingress port. It's difficult to take
1879 advantage of OFPP_IN_PORT in this situation. To help, Open vSwitch
1880 provides, as an OpenFlow extension, the ability to modify the
1881 in_port field. Whatever value is currently in the in_port field is
1882 the port to which outputs will be dropped, as well as the
1883 destination for OFPP_IN_PORT. This means that the following will
1884 reliably output to port 2 or to ports 2 through 6, respectively:
1886 ovs-ofctl add-flow br0 in_port=2,actions=load:0->NXM_OF_IN_PORT[],2
1887 ovs-ofctl add-flow br0 actions=load:0->NXM_OF_IN_PORT[],2,3,4,5,6
1889 If the input port is important, then one may save and restore it on
1892 ovs-ofctl add-flow br0 actions=push:NXM_OF_IN_PORT[],\
1893 load:0->NXM_OF_IN_PORT[],\
1895 pop:NXM_OF_IN_PORT[]
1897 ### Q: My bridge br0 has host 192.168.0.1 on port 1 and host 192.168.0.2
1898 on port 2. I set up flows to forward only traffic destined to the
1899 other host and drop other traffic, like this:
1901 priority=5,in_port=1,ip,nw_dst=192.168.0.2,actions=2
1902 priority=5,in_port=2,ip,nw_dst=192.168.0.1,actions=1
1903 priority=0,actions=drop
1905 But it doesn't work--I don't get any connectivity when I do this.
1908 A: These flows drop the ARP packets that IP hosts use to establish IP
1909 connectivity over Ethernet. To solve the problem, add flows to
1910 allow ARP to pass between the hosts:
1912 priority=5,in_port=1,arp,actions=2
1913 priority=5,in_port=2,arp,actions=1
1915 This issue can manifest other ways, too. The following flows that
1916 match on Ethernet addresses instead of IP addresses will also drop
1917 ARP packets, because ARP requests are broadcast instead of being
1918 directed to a specific host:
1920 priority=5,in_port=1,dl_dst=54:00:00:00:00:02,actions=2
1921 priority=5,in_port=2,dl_dst=54:00:00:00:00:01,actions=1
1922 priority=0,actions=drop
1924 The solution already described above will also work in this case.
1925 It may be better to add flows to allow all multicast and broadcast
1928 priority=5,in_port=1,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=2
1929 priority=5,in_port=2,dl_dst=01:00:00:00:00:00/01:00:00:00:00:00,actions=1
1931 ### Q: My bridge disconnects from my controller on add-port/del-port.
1933 A: Reconfiguring your bridge can change your bridge's datapath-id because
1934 Open vSwitch generates datapath-id from the MAC address of one of its ports.
1935 In that case, Open vSwitch disconnects from controllers because there's
1936 no graceful way to notify controllers about the change of datapath-id.
1938 To avoid the behaviour, you can configure datapath-id manually.
1940 ovs-vsctl set bridge br0 other-config:datapath-id=0123456789abcdef
1942 ### Q: My controller is getting errors about "buffers". What's going on?
1944 A: When a switch sends a packet to an OpenFlow controller using a
1945 "packet-in" message, it can also keep a copy of that packet in a
1946 "buffer", identified by a 32-bit integer "buffer_id". There are
1947 two advantages to buffering. First, when the controller wants to
1948 tell the switch to do something with the buffered packet (with a
1949 "packet-out" OpenFlow request), it does not need to send another
1950 copy of the packet back across the OpenFlow connection, which
1951 reduces the bandwidth cost of the connection and improves latency.
1952 This enables the second advantage: the switch can optionally send
1953 only the first part of the packet to the controller (assuming that
1954 the switch only needs to look at the first few bytes of the
1955 packet), further reducing bandwidth and improving latency.
1957 However, buffering introduces some issues of its own. First, any
1958 switch has limited resources, so if the controller does not use a
1959 buffered packet, the switch has to decide how long to keep it
1960 buffered. When many packets are sent to a controller and buffered,
1961 Open vSwitch can discard buffered packets that the controller has
1962 not used after as little as 5 seconds. This means that
1963 controllers, if they make use of packet buffering, should use the
1964 buffered packets promptly. (This includes sending a "packet-out"
1965 with no actions if the controller does not want to do anything with
1966 a buffered packet, to clear the packet buffer and effectively
1969 Second, packet buffers are one-time-use, meaning that a controller
1970 cannot use a single packet buffer in two or more "packet-out"
1971 commands. Open vSwitch will respond with an error to the second
1972 and subsequent "packet-out"s in such a case.
1974 Finally, a common error early in controller development is to try
1975 to use buffer_id 0 in a "packet-out" message as if 0 represented
1976 "no buffered packet". This is incorrect usage: the buffer_id with
1977 this meaning is actually 0xffffffff.
1979 ovs-vswitchd(8) describes some details of Open vSwitch packet
1980 buffering that the OpenFlow specification requires implementations
1983 ### Q: How does OVS divide flows among buckets in an OpenFlow "select" group?
1985 A: In Open vSwitch 2.3 and earlier, Open vSwitch used the destination
1986 Ethernet address to choose a bucket in a select group.
1988 Open vSwitch 2.4 and later by default hashes the source and
1989 destination Ethernet address, VLAN ID, Ethernet type, IPv4/v6
1990 source and destination address and protocol, and for TCP and SCTP
1991 only, the source and destination ports. The hash is "symmetric",
1992 meaning that exchanging source and destination addresses does not
1993 change the bucket selection.
1995 Select groups in Open vSwitch 2.4 and later can be configured to
1996 use a different hash function, using a Netronome extension to the
1997 OpenFlow 1.5+ group_mod message. For more information, see
1998 Documentation/group-selection-method-property.txt in the Open
1999 vSwitch source tree. (OpenFlow 1.5 support in Open vSwitch is still
2006 ### Q: How do I implement a new OpenFlow message?
2008 A: Add your new message to "enum ofpraw" and "enum ofptype" in
2009 lib/ofp-msgs.h, following the existing pattern. Then recompile and
2010 fix all of the new warnings, implementing new functionality for the
2011 new message as needed. (If you configure with --enable-Werror, as
2012 described in [INSTALL.md], then it is impossible to miss any warnings.)
2014 If you need to add an OpenFlow vendor extension message for a
2015 vendor that doesn't yet have any extension messages, then you will
2016 also need to edit build-aux/extract-ofp-msgs.
2018 ### Q: How do I add support for a new field or header?
2020 A: Add new members for your field to "struct flow" in lib/flow.h, and
2021 add new enumerations for your new field to "enum mf_field_id" in
2022 lib/meta-flow.h, following the existing pattern. Also, add support
2023 to miniflow_extract() in lib/flow.c for extracting your new field
2024 from a packet into struct miniflow, and to nx_put_raw() in
2025 lib/nx-match.c to output your new field in OXM matches. Then
2026 recompile and fix all of the new warnings, implementing new
2027 functionality for the new field or header as needed. (If you
2028 configure with --enable-Werror, as described in [INSTALL.md], then
2029 it is impossible to miss any warnings.)
2031 If you want kernel datapath support for your new field, you also
2032 need to modify the kernel module for the operating systems you are
2033 interested in. This isn't mandatory, since fields understood only
2034 by userspace work too (with a performance penalty), so it's
2035 reasonable to start development without it. If you implement
2036 kernel module support for Linux, then the Linux kernel "netdev"
2037 mailing list is the place to submit that support first; please read
2038 up on the Linux kernel development process separately. The Windows
2039 datapath kernel module support, on the other hand, is maintained
2040 within the OVS tree, so patches for that can go directly to
2043 ### Q: How do I add support for a new OpenFlow action?
2045 A: Add your new action to "enum ofp_raw_action_type" in
2046 lib/ofp-actions.c, following the existing pattern. Then recompile
2047 and fix all of the new warnings, implementing new functionality for
2048 the new action as needed. (If you configure with --enable-Werror,
2049 as described in [INSTALL.md], then it is impossible to miss any
2052 If you need to add an OpenFlow vendor extension action for a vendor
2053 that doesn't yet have any extension actions, then you will also
2054 need to edit build-aux/extract-ofp-actions.
2060 bugs@openvswitch.org
2061 http://openvswitch.org/
2063 [PORTING.md]:PORTING.md
2064 [WHY-OVS.md]:WHY-OVS.md
2065 [INSTALL.md]:INSTALL.md
2066 [OPENFLOW-1.1+.md]:OPENFLOW-1.1+.md
2067 [INSTALL.DPDK.md]:INSTALL.DPDK.md
projects / cascardo / ovs.git / blob