1 How to Install Open vSwitch on Linux, FreeBSD and NetBSD
2 ========================================================
4 This document describes how to build and install Open vSwitch on a
5 generic Linux, FreeBSD, or NetBSD host. For specifics around installation
6 on a specific platform, please see one of these files:
11 - [INSTALL.XenServer.md]
18 To compile the userspace programs in the Open vSwitch distribution,
19 you will need the following software:
23 - A C compiler, such as:
27 * Clang. Clang 3.4 and later provide useful static semantic
28 analysis and thread-safety checks. For Ubuntu, there are
29 nightly built packages available on clang's website.
31 While OVS may be compatible with other compilers, optimal
32 support for atomic operations may be missing, making OVS very
33 slow (see lib/ovs-atomic.h).
35 - libssl, from OpenSSL, is optional but recommended if you plan to
36 connect the Open vSwitch to an OpenFlow controller. libssl is
37 required to establish confidentiality and authenticity in the
38 connections from an Open vSwitch to an OpenFlow controller. If
39 libssl is installed, then Open vSwitch will automatically build
42 - Python 2.x, for x >= 4.
44 On Linux, you may choose to compile the kernel module that comes with
45 the Open vSwitch distribution or to use the kernel module built into
46 the Linux kernel (version 3.3 or later). See the [FAQ.md] question
47 "What features are not available in the Open vSwitch kernel datapath that
48 ships as part of the upstream Linux kernel?" for more information on
49 this trade-off. You may also use the userspace-only implementation,
50 at some cost in features and performance (see [INSTALL.userspace.md]
51 for details). To compile the kernel module on Linux, you must also
52 install the following:
54 - A supported Linux kernel version. Please refer to [README.md] for a
55 list of supported versions.
57 The Open vSwitch datapath requires bridging support
58 (CONFIG_BRIDGE) to be built as a kernel module. (This is common
59 in kernels provided by Linux distributions.) The bridge module
60 must not be loaded or in use. If the bridge module is running
61 (check with "lsmod | grep bridge"), you must remove it ("rmmod
62 bridge") before starting the datapath.
64 For optional support of ingress policing, you must enable kernel
65 configuration options NET_CLS_BASIC, NET_SCH_INGRESS, and
66 NET_ACT_POLICE, either built-in or as modules. (NET_CLS_POLICE is
67 obsolete and not needed.)
69 To use GRE tunneling on Linux 2.6.37 or newer, kernel support
70 for GRE demultiplexing (CONFIG_NET_IPGRE_DEMUX) must be compiled
71 in or available as a module. Also, on kernels before 3.11, the
72 ip_gre module, for GRE tunnels over IP (NET_IPGRE), must not be
73 loaded or compiled in.
75 To configure HTB or HFSC quality of service with Open vSwitch,
76 you must enable the respective configuration options.
78 To use Open vSwitch support for TAP devices, you must enable
81 - To build a kernel module, you need the same version of GCC that
82 was used to build that kernel.
84 - A kernel build directory corresponding to the Linux kernel image
85 the module is to run on. Under Debian and Ubuntu, for example,
86 each linux-image package containing a kernel binary has a
87 corresponding linux-headers package with the required build
90 If you are working from a Git tree or snapshot (instead of from a
91 distribution tarball), or if you modify the Open vSwitch build system
92 or the database schema, you will also need the following software:
94 - Autoconf version 2.63 or later.
96 - Automake version 1.10 or later.
98 - libtool version 2.4 or later. (Older versions might work too.)
100 To run the unit tests, you also need:
102 - Perl. Version 5.10.1 is known to work. Earlier versions should
105 The ovs-vswitchd.conf.db(5) manpage will include an E-R diagram, in
106 formats other than plain text, only if you have the following:
108 - "dot" from graphviz (http://www.graphviz.org/).
110 - Perl. Version 5.10.1 is known to work. Earlier versions should
113 - Python 2.x, for x >= 4.
115 If you are going to extensively modify Open vSwitch, please consider
116 installing the following to obtain better warnings:
118 - "sparse" version 0.4.4 or later
119 (http://www.kernel.org/pub/software/devel/sparse/dist/).
123 - clang, version 3.4 or later
125 Also, you may find the ovs-dev script found in utilities/ovs-dev.py useful.
127 Installation Requirements
128 -------------------------
130 The machine on which Open vSwitch is to be installed must have the
133 - libc compatible with the libc used for build.
135 - libssl compatible with the libssl used for build, if OpenSSL was
138 - On Linux, the same kernel version configured as part of the build.
140 - For optional support of ingress policing on Linux, the "tc" program
141 from iproute2 (part of all major distributions and available at
142 http://www.linux-foundation.org/en/Net:Iproute2).
144 On Linux you should ensure that /dev/urandom exists. To support TAP
145 devices, you must also ensure that /dev/net/tun exists.
147 Building and Installing Open vSwitch for Linux, FreeBSD or NetBSD
148 =================================================================
150 Once you have installed all the prerequisites listed above in the Base
151 Prerequisites section, follow the procedure below to build.
153 1. If you pulled the sources directly from an Open vSwitch Git tree,
154 run boot.sh in the top source directory:
158 2. Configure the package by running the configure script. You can
159 usually invoke configure without any arguments. For example:
163 By default all files are installed under /usr/local. If you want
164 to install into, e.g., /usr and /var instead of /usr/local and
165 /usr/local/var, add options as shown here:
167 `% ./configure --prefix=/usr --localstatedir=/var`
169 To use a specific C compiler for compiling Open vSwitch user
170 programs, also specify it on the configure command line, like so:
172 `% ./configure CC=gcc-4.2`
174 To use 'clang' compiler:
176 `% ./configure CC=clang`
178 To build the Linux kernel module, so that you can run the
179 kernel-based switch, pass the location of the kernel build
180 directory on --with-linux. For example, to build for a running
183 `% ./configure --with-linux=/lib/modules/`uname -r`/build`
185 If --with-linux requests building for an unsupported version of
186 Linux, then "configure" will fail with an error message. Please
187 refer to the [FAQ.md] for advice in that case.
189 If you wish to build the kernel module for an architecture other
190 than the architecture of the machine used for the build, you may
191 specify the kernel architecture string using the KARCH variable
192 when invoking the configure script. For example, to build for MIPS
195 `% ./configure --with-linux=/path/to/linux KARCH=mips`
197 If you plan to do much Open vSwitch development, you might want to
198 add --enable-Werror, which adds the -Werror option to the compiler
199 command line, turning warnings into errors. That makes it
200 impossible to miss warnings generated by the build.
202 To build with gcov code coverage support, add --enable-coverage,
205 `% ./configure --enable-coverage`
207 The configure script accepts a number of other options and honors
208 additional environment variables. For a full list, invoke
209 configure with the --help option.
211 You can also run configure from a separate build directory. This
212 is helpful if you want to build Open vSwitch in more than one way
213 from a single source directory, e.g. to try out both GCC and Clang
214 builds, or to build kernel modules for more than one Linux version.
217 `% mkdir _gcc && (cd _gcc && ../configure CC=gcc)`
218 `% mkdir _clang && (cd _clang && ../configure CC=clang)`
220 3. Run GNU make in the build directory, e.g.:
224 or if GNU make is installed as "gmake":
228 If you used a separate build directory, run make or gmake from that
234 For improved warnings if you installed "sparse" (see
235 "Prerequisites"), add C=1 to the command line.
237 4. Consider running the testsuite. Refer to "Running the Testsuite"
238 below, for instructions.
240 5. Become root by running "su" or another program.
242 6. Run "make install" to install the executables and manpages into the
243 running system, by default under /usr/local.
245 7. If you built kernel modules, you may install and load them, e.g.:
247 `% make modules_install`
248 `% /sbin/modprobe openvswitch`
250 To verify that the modules have been loaded, run "/sbin/lsmod" and
251 check that openvswitch is listed.
253 If the `modprobe` operation fails, look at the last few kernel log
254 messages (e.g. with `dmesg | tail`):
256 - The message "openvswitch: exports duplicate symbol
257 br_should_route_hook (owned by bridge)" means that the bridge
258 module is loaded. Run `/sbin/rmmod bridge` to remove it.
260 If `/sbin/rmmod bridge` fails with "ERROR: Module bridge does
261 not exist in /proc/modules", then the bridge is compiled into
262 the kernel, rather than as a module. Open vSwitch does not
263 support this configuration (see "Build Requirements", above).
265 - The message "openvswitch: exports duplicate symbol
266 dp_ioctl_hook (owned by ofdatapath)" means that the ofdatapath
267 module from the OpenFlow reference implementation is loaded.
268 Run `/sbin/rmmod ofdatapath` to remove it. (You might have to
269 delete any existing datapaths beforehand, using the "dpctl"
270 program included with the OpenFlow reference implementation.
271 "ovs-dpctl" will not work.)
273 - Otherwise, the most likely problem is that Open vSwitch was
274 built for a kernel different from the one into which you are
275 trying to load it. Run `modinfo` on openvswitch.ko and on
276 a module built for the running kernel, e.g.:
279 % /sbin/modinfo openvswitch.ko
280 % /sbin/modinfo /lib/modules/`uname -r`/kernel/net/bridge/bridge.ko
283 Compare the "vermagic" lines output by the two commands. If
284 they differ, then Open vSwitch was built for the wrong kernel.
286 - If you decide to report a bug or ask a question related to
287 module loading, please include the output from the `dmesg` and
288 `modinfo` commands mentioned above.
290 There is an optional module parameter to openvswitch.ko called
291 vlan_tso that enables TCP segmentation offload over VLANs on NICs
292 that support it. Many drivers do not expose support for TSO on VLANs
293 in a way that Open vSwitch can use but there is no way to detect
294 whether this is the case. If you know that your particular driver can
295 handle it (for example by testing sending large TCP packets over VLANs)
296 then passing in a value of 1 may improve performance. Modules built for
297 Linux kernels 2.6.37 and later, as well as specially patched versions
298 of earlier kernels, do not need this and do not have this parameter. If
299 you do not understand what this means or do not know if your driver
300 will work, do not set this.
302 8. Initialize the configuration database using ovsdb-tool, e.g.:
304 `% mkdir -p /usr/local/etc/openvswitch`
305 `% ovsdb-tool create /usr/local/etc/openvswitch/conf.db vswitchd/vswitch.ovsschema`
310 Before starting ovs-vswitchd itself, you need to start its
311 configuration database, ovsdb-server. Each machine on which Open
312 vSwitch is installed should run its own copy of ovsdb-server.
313 Configure it to use the database you created during installation (as
314 explained above), to listen on a Unix domain socket, to connect to any
315 managers specified in the database itself, and to use the SSL
316 configuration in the database:
318 % ovsdb-server --remote=punix:/usr/local/var/run/openvswitch/db.sock \
319 --remote=db:Open_vSwitch,Open_vSwitch,manager_options \
320 --private-key=db:Open_vSwitch,SSL,private_key \
321 --certificate=db:Open_vSwitch,SSL,certificate \
322 --bootstrap-ca-cert=db:Open_vSwitch,SSL,ca_cert \
325 (If you built Open vSwitch without SSL support, then omit
326 --private-key, --certificate, and --bootstrap-ca-cert.)
328 Then initialize the database using ovs-vsctl. This is only
329 necessary the first time after you create the database with
330 ovsdb-tool (but running it at any time is harmless):
332 % ovs-vsctl --no-wait init
334 Then start the main Open vSwitch daemon, telling it to connect to the
335 same Unix domain socket:
337 % ovs-vswitchd --pidfile --detach
339 Now you may use ovs-vsctl to set up bridges and other Open vSwitch
340 features. For example, to create a bridge named br0 and add ports
341 eth0 and vif1.0 to it:
343 % ovs-vsctl add-br br0
344 % ovs-vsctl add-port br0 eth0
345 % ovs-vsctl add-port br0 vif1.0
347 Please refer to ovs-vsctl(8) for more details.
352 When you upgrade Open vSwitch from one version to another, you should
353 also upgrade the database schema:
355 1. Stop the Open vSwitch daemons, e.g.:
358 % kill `cd /usr/local/var/run/openvswitch && cat ovsdb-server.pid ovs-vswitchd.pid`
361 2. Install the new Open vSwitch release.
363 3. Upgrade the database, in one of the following two ways:
365 - If there is no important data in your database, then you may
366 delete the database file and recreate it with ovsdb-tool,
367 following the instructions under "Building and Installing Open
368 vSwitch for Linux, FreeBSD or NetBSD".
370 - If you want to preserve the contents of your database, back it
371 up first, then use "ovsdb-tool convert" to upgrade it, e.g.:
373 `% ovsdb-tool convert /usr/local/etc/openvswitch/conf.db vswitchd/vswitch.ovsschema`
375 4. Start the Open vSwitch daemons as described under "Building and
376 Installing Open vSwitch for Linux, FreeBSD or NetBSD" above.
380 Upgrading Open vSwitch from one version to the next version with minimum
381 disruption of traffic going through the system that is using that Open vSwitch
382 needs some considerations:
384 1. If the upgrade only involves upgrading the userspace utilities and daemons
385 of Open vSwitch, make sure that the new userspace version is compatible with
386 the previously loaded kernel module.
388 2. An upgrade of userspace daemons means that they have to be restarted.
389 Restarting the daemons means that the OpenFlow flows in the ovs-vswitchd daemon
390 will be lost. One way to restore the flows is to let the controller
391 re-populate it. Another way is to save the previous flows using a utility
392 like ovs-ofctl and then re-add them after the restart. Restoring the old flows
393 is accurate only if the new Open vSwitch interfaces retain the old 'ofport'
396 3. When the new userspace daemons get restarted, they automatically flush
397 the old flows setup in the kernel. This can be expensive if there are hundreds
398 of new flows that are entering the kernel but userspace daemons are busy
399 setting up new userspace flows from either the controller or an utility like
400 ovs-ofctl. Open vSwitch database provides an option to solve this problem
401 through the other_config:flow-restore-wait column of the Open_vSwitch table.
402 Refer to the ovs-vswitchd.conf.db(5) manpage for details.
404 4. If the upgrade also involves upgrading the kernel module, the old kernel
405 module needs to be unloaded and the new kernel module should be loaded. This
406 means that the kernel network devices belonging to Open vSwitch is recreated
407 and the kernel flows are lost. The downtime of the traffic can be reduced
408 if the userspace daemons are restarted immediately and the userspace flows
409 are restored as soon as possible.
411 The ovs-ctl utility's "restart" function only restarts the userspace daemons,
412 makes sure that the 'ofport' values remain consistent across restarts, restores
413 userspace flows using the ovs-ofctl utility and also uses the
414 other_config:flow-restore-wait column to keep the traffic downtime to the
415 minimum. The ovs-ctl utility's "force-reload-kmod" function does all of the
416 above, but also replaces the old kernel module with the new one. Open vSwitch
417 startup scripts for Debian, XenServer and RHEL use ovs-ctl's functions and it
418 is recommended that these functions be used for other software platforms too.
423 This section describe Open vSwitch's built-in support for various test
424 suites. You must configure and build Open vSwitch (steps 1 through 3
425 in "Building and Installing Open vSwitch for Linux, FreeBSD or NetBSD"
426 above) before you run the tests described here. You do not need to
427 install Open vSwitch or to build or load the kernel module to run
428 these test suites. You do not need supervisor privilege to run these
434 Open vSwitch includes a suite of self-tests. Before you submit patches
435 upstream, we advise that you run the tests and ensure that they pass.
436 If you add new features to Open vSwitch, then adding tests for those
437 features will ensure your features don't break as developers modify
438 other areas of Open vSwitch.
440 Refer to "Testsuites" above for prerequisites.
442 To run all the unit tests in Open vSwitch, one at a time:
444 This takes under 5 minutes on a modern desktop system.
446 To run all the unit tests in Open vSwitch, up to 8 in parallel:
447 `make check TESTSUITEFLAGS=-j8`
448 This takes under a minute on a modern 4-core desktop system.
450 To see a list of all the available tests, run:
451 `make check TESTSUITEFLAGS=--list`
453 To run only a subset of tests, e.g. test 123 and tests 477 through 484:
454 `make check TESTSUITEFLAGS='123 477-484'`
455 (Tests do not have inter-dependencies, so you may run any subset.)
457 To run tests matching a keyword, e.g. "ovsdb":
458 `make check TESTSUITEFLAGS='-k ovsdb'`
460 To see a complete list of test options:
461 `make check TESTSUITEFLAGS=--help`
463 The results of a testing run are reported in tests/testsuite.log.
464 Please report test failures as bugs and include the testsuite.log in
467 If you have "valgrind" installed, then you can also run the testsuite
468 under valgrind by using "make check-valgrind" in place of "make
469 check". All the same options are available via TESTSUITEFLAGS. When
470 you do this, the "valgrind" results for test `<N>` are reported in files
471 named `tests/testsuite.dir/<N>/valgrind.*`. You may find that the
472 valgrind results are easier to interpret if you put "-q" in
473 ~/.valgrindrc, since that reduces the amount of output.
475 Sometimes a few tests may fail on some runs but not others. This is
476 usually a bug in the testsuite, not a bug in Open vSwitch itself. If
477 you find that a test fails intermittently, please report it, since the
478 developers may not have noticed.
483 OFTest is an OpenFlow protocol testing suite. Open vSwitch includes a
484 Makefile target to run OFTest with Open vSwitch in "dummy mode". In
485 this mode of testing, no packets travel across physical or virtual
486 networks. Instead, Unix domain sockets stand in as simulated
487 networks. This simulation is imperfect, but it is much easier to set
488 up, does not require extra physical or virtual hardware, and does not
489 require supervisor privileges.
491 To run OFTest with Open vSwitch, first read and follow the
492 instructions under "Testsuites" above. Second, obtain a copy of
493 OFTest and install its prerequisites. You need a copy of OFTest that
494 includes commit 406614846c5 (make ovs-dummy platform work again).
495 This commit was merged into the OFTest repository on Feb 1, 2013, so
496 any copy of OFTest more recent than that should work. Testing OVS in
497 dummy mode does not require root privilege, so you may ignore that
500 Optionally, add the top-level OFTest directory (containing the "oft"
501 program) to your $PATH. This slightly simplifies running OFTest later.
503 To run OFTest in dummy mode, run the following command from your Open
504 vSwitch build directory:
505 `make check-oftest OFT=<oft-binary>`
506 where `<oft-binary>` is the absolute path to the "oft" program in
509 If you added "oft" to your $PATH, you may omit the OFT variable
512 By default, "check-oftest" passes "oft" just enough options to enable
513 dummy mode. You can use OFTFLAGS to pass additional options. For
514 example, to run just the basic.Echo test instead of all tests (the
515 default) and enable verbose logging:
516 `make check-oftest OFT=<oft-binary> OFTFLAGS='--verbose -T basic.Echo'`
518 If you use OFTest that does not include commit 4d1f3eb2c792 (oft:
519 change default port to 6653), merged into the OFTest repository in
520 October 2013, then you need to add an option to use the IETF-assigned
522 `make check-oftest OFT=<oft-binary> OFTFLAGS='--port=6653'`
524 Please interpret OFTest results cautiously. Open vSwitch can fail a
525 given test in OFTest for many reasons, including bugs in Open vSwitch,
526 bugs in OFTest, bugs in the "dummy mode" integration, and differing
527 interpretations of the OpenFlow standard and other standards.
529 Open vSwitch has not been validated against OFTest. Please do report
530 test failures that you believe to represent bugs in Open vSwitch.
531 Include the precise versions of Open vSwitch and OFTest in your bug
532 report, plus any other information needed to reproduce the problem.
537 Ryu is an OpenFlow controller written in Python that includes an
538 extensive OpenFlow testsuite. Open vSwitch includes a Makefile target
539 to run Ryu in "dummy mode". See "OFTest" above for an explanation of
542 To run Ryu tests with Open vSwitch, first read and follow the
543 instructions under "Testsuites" above. Second, obtain a copy of Ryu,
544 install its prerequisites, and build it. You do not need to install
545 Ryu (some of the tests do not get installed, so it does not help).
547 To run Ryu tests, run the following command from your Open vSwitch
549 `make check-ryu RYUDIR=<ryu-source-dir>`
550 where `<ryu-source-dir>` is the absolute path to the root of the Ryu
551 source distribution. The default `<ryu-source-dir>` is `$srcdir/../ryu`
552 where $srcdir is your Open vSwitch source directory, so if this
553 default is correct then you make simply run `make check-ryu`.
555 Open vSwitch has not been validated against Ryu. Please do report
556 test failures that you believe to represent bugs in Open vSwitch.
557 Include the precise versions of Open vSwitch and Ryu in your bug
558 report, plus any other information needed to reproduce the problem.
560 Continuous Integration with Travis-CI
561 -------------------------------------
563 A .travis.yml file is provided to automatically build Open vSwitch with
564 various build configurations and run the testsuite using travis-ci.
565 Builds will be performed with gcc, sparse and clang with the -Werror
566 compiler flag included, therefore the build will fail if a new warning
569 The CI build is triggered via git push (regardless of the specific
570 branch) or pull request against any Open vSwitch GitHub repository that
571 is linked to travis-ci.
573 Instructions to setup travis-ci for your GitHub repository:
575 1. Go to http://travis-ci.org/ and sign in using your GitHub ID.
576 2. Go to the "Repositories" tab and enable the ovs repository. You
577 may disable builds for pushes or pull requests.
578 3. In order to avoid forks sending build failures to the upstream
579 mailing list, the notification email recipient is encrypted. If you
580 want to receive email notification for build failures, replace the
581 the encrypted string:
582 3.1) Install the travis-ci CLI (Requires ruby >=2.0):
584 3.2) In your Open vSwitch repository:
585 travis encrypt mylist@mydomain.org
586 3.3) Add/replace the notifications section in .travis.yml and fill
587 in the secure string as returned by travis encrypt:
594 (You may remove/omit the notifications section to fall back to
595 default notification behaviour which is to send an email directly
596 to the author and committer of the failing commit. Note that the
597 email is only sent if the author/committer have commit rights for
598 the particular GitHub repository).
600 4. Pushing a commit to the repository which breaks the build or the
601 testsuite will now trigger a email sent to mylist@mydomain.org
606 Please report problems to bugs@openvswitch.org.
608 [README.md]:README.md
609 [INSTALL.Debian.md]:INSTALL.Debian.md
610 [INSTALL.Fedora.md]:INSTALL.Fedora.md
611 [INSTALL.RHEL.md]:INSTALL.RHEL.md
612 [INSTALL.XenServer.md]:INSTALL.XenServer.md
613 [INSTALL.NetBSD.md]:INSTALL.NetBSD.md
614 [INSTALL.DPDK.md]:INSTALL.DPDK.md
615 [INSTALL.userspace.md]:INSTALL.userspace.md