1 .. This work is licensed under a Creative Commons Attribution 4.0 International License.
2 .. http://creativecommons.org/licenses/by/4.0
3 .. (c) OPNFV, Intel Corporation, AT&T and others.
5 vSwitchPerf test suites userguide
6 ---------------------------------
11 VSPERF requires a traffic generators to run tests, automated traffic gen
12 support in VSPERF includes:
14 - IXIA traffic generator (IxNetwork hardware) and a machine that runs the IXIA
16 - Spirent traffic generator (TestCenter hardware chassis or TestCenter virtual
17 in a VM) and a VM to run the Spirent Virtual Deployment Service image,
18 formerly known as "Spirent LabServer".
19 - Xena Network traffic generator (Xena hardware chassis) that houses the Xena
20 Traffic generator modules.
21 - Moongen software traffic generator. Requires a separate machine running
22 moongen to execute packet generation.
23 - T-Rex software traffic generator. Requires a separate machine running T-Rex
24 Server to execute packet generation.
26 If you want to use another traffic generator, please select the :ref:`trafficgen-dummy`
32 To see the supported Operating Systems, vSwitches and system requirements,
33 please follow the `installation instructions <vsperf-installation>`.
35 Traffic Generator Setup
36 ^^^^^^^^^^^^^^^^^^^^^^^
38 Follow the `Traffic generator instructions <trafficgen-installation>` to
39 install and configure a suitable traffic generator.
41 Cloning and building src dependencies
42 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
44 In order to run VSPERF, you will need to download DPDK and OVS. You can
45 do this manually and build them in a preferred location, OR you could
46 use vswitchperf/src. The vswitchperf/src directory contains makefiles
47 that will allow you to clone and build the libraries that VSPERF depends
48 on, such as DPDK and OVS. To clone and build simply:
50 .. code-block:: console
55 VSPERF can be used with stock OVS (without DPDK support). When build
56 is finished, the libraries are stored in src_vanilla directory.
58 The 'make' builds all options in src:
61 * OVS with vhost_user as the guest access method (with DPDK support)
63 The vhost_user build will reside in src/ovs/
64 The Vanilla OVS build will reside in vswitchperf/src_vanilla
66 To delete a src subdirectory and its contents to allow you to re-clone simply
69 .. code-block:: console
73 Configure the ``./conf/10_custom.conf`` file
74 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
76 The ``10_custom.conf`` file is the configuration file that overrides
77 default configurations in all the other configuration files in ``./conf``
78 The supplied ``10_custom.conf`` file **MUST** be modified, as it contains
79 configuration items for which there are no reasonable default values.
81 The configuration items that can be added is not limited to the initial
82 contents. Any configuration item mentioned in any .conf file in
83 ``./conf`` directory can be added and that item will be overridden by
84 the custom configuration value.
86 Further details about configuration files evaluation and special behaviour
87 of options with ``GUEST_`` prefix could be found at :ref:`design document
88 <design-configuration>`.
90 Using a custom settings file
91 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
93 If your ``10_custom.conf`` doesn't reside in the ``./conf`` directory
94 of if you want to use an alternative configuration file, the file can
95 be passed to ``vsperf`` via the ``--conf-file`` argument.
97 .. code-block:: console
99 $ ./vsperf --conf-file <path_to_custom_conf> ...
101 Note that configuration passed in via the environment (``--load-env``)
102 or via another command line argument will override both the default and
103 your custom configuration files. This "priority hierarchy" can be
104 described like so (1 = max priority):
106 1. Testcase definition section ``Parameters``
107 2. Command line arguments
108 3. Environment variables
109 4. Configuration file(s)
111 Further details about configuration files evaluation and special behaviour
112 of options with ``GUEST_`` prefix could be found at :ref:`design document
113 <design-configuration>`.
115 .. _overriding-parameters-documentation:
117 Referencing parameter values
118 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
120 It is possible to use a special macro ``#PARAM()`` to refer to the value of
121 another configuration parameter. This reference is evaluated during
122 access of the parameter value (by ``settings.getValue()`` call), so it
123 can refer to parameters created during VSPERF runtime, e.g. NICS dictionary.
124 It can be used to reflect DUT HW details in the testcase definition.
136 # set destination MAC to the MAC of the first
137 # interface from WHITELIST_NICS list
138 'dstmac' : '#PARAM(NICS[0]["mac"])',
143 Overriding values defined in configuration files
144 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
146 The configuration items can be overridden by command line argument
147 ``--test-params``. In this case, the configuration items and
148 their values should be passed in form of ``item=value`` and separated
155 $ ./vsperf --test-params "TRAFFICGEN_DURATION=10;TRAFFICGEN_PKT_SIZES=(128,);" \
156 "GUEST_LOOPBACK=['testpmd','l2fwd']" pvvp_tput
158 The second option is to override configuration items by ``Parameters`` section
159 of the test case definition. The configuration items can be added into ``Parameters``
160 dictionary with their new values. These values will override values defined in
161 configuration files or specified by ``--test-params`` command line argument.
167 "Parameters" : {'TRAFFICGEN_PKT_SIZES' : (128,),
168 'TRAFFICGEN_DURATION' : 10,
169 'GUEST_LOOPBACK' : ['testpmd','l2fwd'],
172 **NOTE:** In both cases, configuration item names and their values must be specified
173 in the same form as they are defined inside configuration files. Parameter names
174 must be specified in uppercase and data types of original and new value must match.
175 Python syntax rules related to data types and structures must be followed.
176 For example, parameter ``TRAFFICGEN_PKT_SIZES`` above is defined as a tuple
177 with a single value ``128``. In this case trailing comma is mandatory, otherwise
178 value can be wrongly interpreted as a number instead of a tuple and vsperf
179 execution would fail. Please check configuration files for default values and their
180 types and use them as a basis for any customized values. In case of any doubt, please
181 check official python documentation related to data structures like tuples, lists
184 **NOTE:** Vsperf execution will terminate with runtime error in case, that unknown
185 parameter name is passed via ``--test-params`` CLI argument or defined in ``Parameters``
186 section of test case definition. It is also forbidden to redefine a value of
187 ``TEST_PARAMS`` configuration item via CLI or ``Parameters`` section.
192 VSPERF uses a VM image called vloop_vnf for looping traffic in the deployment
193 scenarios involving VMs. The image can be downloaded from
194 `<http://artifacts.opnfv.org/>`__.
196 Please see the installation instructions for information on :ref:`vloop-vnf`
204 A Kernel Module that provides OSI Layer 2 Ipv4 termination or forwarding with
205 support for Destination Network Address Translation (DNAT) for both the MAC and
206 IP addresses. l2fwd can be found in <vswitchperf_dir>/src/l2fwd
211 All examples inside these docs assume, that user is inside the VSPERF
212 directory. VSPERF can be executed from any directory.
214 Before running any tests make sure you have root permissions by adding
215 the following line to /etc/sudoers:
217 .. code-block:: console
219 username ALL=(ALL) NOPASSWD: ALL
221 username in the example above should be replaced with a real username.
223 To list the available tests:
225 .. code-block:: console
229 To run a single test:
231 .. code-block:: console
235 Where $TESTNAME is the name of the vsperf test you would like to run.
237 To run a group of tests, for example all tests with a name containing
240 .. code-block:: console
242 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf --tests="RFC2544"
246 .. code-block:: console
248 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
250 Some tests allow for configurable parameters, including test duration
251 (in seconds) as well as packet sizes (in bytes).
255 $ ./vsperf --conf-file user_settings.py \
256 --tests RFC2544Tput \
257 --test-params "TRAFFICGEN_DURATION=10;TRAFFICGEN_PKT_SIZES=(128,)"
259 For all available options, check out the help dialog:
261 .. code-block:: console
265 Executing Vanilla OVS tests
266 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
268 1. If needed, recompile src for all OVS variants
270 .. code-block:: console
276 2. Update your ``10_custom.conf`` file to use Vanilla OVS:
278 .. code-block:: python
280 VSWITCH = 'OvsVanilla'
284 .. code-block:: console
286 $ ./vsperf --conf-file=<path_to_custom_conf>
288 Please note if you don't want to configure Vanilla OVS through the
289 configuration file, you can pass it as a CLI argument.
291 .. code-block:: console
293 $ ./vsperf --vswitch OvsVanilla
296 Executing tests with VMs
297 ^^^^^^^^^^^^^^^^^^^^^^^^
299 To run tests using vhost-user as guest access method:
301 1. Set VSWITCH and VNF of your settings file to:
303 .. code-block:: python
305 VSWITCH = 'OvsDpdkVhost'
306 VNF = 'QemuDpdkVhost'
308 2. If needed, recompile src for all OVS variants
310 .. code-block:: console
318 .. code-block:: console
320 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
322 **NOTE:** By default vSwitch is acting as a server for dpdk vhost-user sockets.
323 In case, that QEMU should be a server for vhost-user sockets, then parameter
324 ``VSWITCH_VHOSTUSER_SERVER_MODE`` should be set to ``False``.
326 Executing tests with VMs using Vanilla OVS
327 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
329 To run tests using Vanilla OVS:
331 1. Set the following variables:
333 .. code-block:: python
335 VSWITCH = 'OvsVanilla'
336 VNF = 'QemuVirtioNet'
338 VANILLA_TGEN_PORT1_IP = n.n.n.n
339 VANILLA_TGEN_PORT1_MAC = nn:nn:nn:nn:nn:nn
341 VANILLA_TGEN_PORT2_IP = n.n.n.n
342 VANILLA_TGEN_PORT2_MAC = nn:nn:nn:nn:nn:nn
344 VANILLA_BRIDGE_IP = n.n.n.n
346 or use ``--test-params`` option
348 .. code-block:: console
350 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
351 --test-params "VANILLA_TGEN_PORT1_IP=n.n.n.n;" \
352 "VANILLA_TGEN_PORT1_MAC=nn:nn:nn:nn:nn:nn;" \
353 "VANILLA_TGEN_PORT2_IP=n.n.n.n;" \
354 "VANILLA_TGEN_PORT2_MAC=nn:nn:nn:nn:nn:nn"
356 2. If needed, recompile src for all OVS variants
358 .. code-block:: console
366 .. code-block:: console
368 $ ./vsperf --conf-file<path_to_custom_conf>/10_custom.conf
375 Currently it is not possible to use standard scenario deployments for execution of
376 tests with VPP. It means, that deployments ``p2p``, ``pvp``, ``pvvp`` and in general any
377 :ref:`pxp-deployment` won't work with VPP. However it is possible to use VPP in
378 :ref:`step-driven-tests`. A basic set of VPP testcases covering ``phy2phy``, ``pvp``
379 and ``pvvp`` tests are already prepared.
381 List of performance tests with VPP support follows:
383 * phy2phy_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
384 * phy2phy_cont_vpp: VPP: Phy2Phy Continuous Stream
385 * phy2phy_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
386 * pvp_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
387 * pvp_cont_vpp: VPP: PVP Continuous Stream
388 * pvp_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
389 * pvvp_tput_vpp: VPP: LTD.Throughput.RFC2544.PacketLossRatio
390 * pvvp_cont_vpp: VPP: PVP Continuous Stream
391 * pvvp_back2back_vpp: VPP: LTD.Throughput.RFC2544.BackToBackFrames
393 In order to execute testcases with VPP it is required to:
395 * install VPP manually, see :ref:`vpp-installation`
396 * configure ``WHITELIST_NICS``, with two physical NICs connected to the traffic generator
397 * configure traffic generator, see :ref:`trafficgen-installation`
399 After that it is possible to execute VPP testcases listed above.
403 .. code-block:: console
405 $ ./vsperf --conf-file=<path_to_custom_conf> phy2phy_tput_vpp
409 Using vfio_pci with DPDK
410 ^^^^^^^^^^^^^^^^^^^^^^^^^
412 To use vfio with DPDK instead of igb_uio add into your custom configuration
413 file the following parameter:
415 .. code-block:: python
417 PATHS['dpdk']['src']['modules'] = ['uio', 'vfio-pci']
420 **NOTE:** In case, that DPDK is installed from binary package, then please
421 set ``PATHS['dpdk']['bin']['modules']`` instead.
423 **NOTE:** Please ensure that Intel VT-d is enabled in BIOS.
425 **NOTE:** Please ensure your boot/grub parameters include
428 **NOTE:** In case of VPP, it is required to explicitly define, that vfio-pci
429 DPDK driver should be used. It means to update dpdk part of VSWITCH_VPP_ARGS
430 dictionary with uio-driver section, e.g. VSWITCH_VPP_ARGS['dpdk'] = 'uio-driver vfio-pci'
432 .. code-block:: console
434 iommu=pt intel_iommu=on
436 To check that IOMMU is enabled on your platform:
438 .. code-block:: console
441 [ 0.000000] Intel-IOMMU: enabled
442 [ 0.139882] dmar: IOMMU 0: reg_base_addr fbffe000 ver 1:0 cap d2078c106f0466 ecap f020de
443 [ 0.139888] dmar: IOMMU 1: reg_base_addr ebffc000 ver 1:0 cap d2078c106f0466 ecap f020de
444 [ 0.139893] IOAPIC id 2 under DRHD base 0xfbffe000 IOMMU 0
445 [ 0.139894] IOAPIC id 0 under DRHD base 0xebffc000 IOMMU 1
446 [ 0.139895] IOAPIC id 1 under DRHD base 0xebffc000 IOMMU 1
447 [ 3.335744] IOMMU: dmar0 using Queued invalidation
448 [ 3.335746] IOMMU: dmar1 using Queued invalidation
456 To use virtual functions of NIC with SRIOV support, use extended form
457 of NIC PCI slot definition:
459 .. code-block:: python
461 WHITELIST_NICS = ['0000:05:00.0|vf0', '0000:05:00.1|vf3']
463 Where 'vf' is an indication of virtual function usage and following
464 number defines a VF to be used. In case that VF usage is detected,
465 then vswitchperf will enable SRIOV support for given card and it will
466 detect PCI slot numbers of selected VFs.
468 So in example above, one VF will be configured for NIC '0000:05:00.0'
469 and four VFs will be configured for NIC '0000:05:00.1'. Vswitchperf
470 will detect PCI addresses of selected VFs and it will use them during
473 At the end of vswitchperf execution, SRIOV support will be disabled.
475 SRIOV support is generic and it can be used in different testing scenarios.
478 * vSwitch tests with DPDK or without DPDK support to verify impact
479 of VF usage on vSwitch performance
480 * tests without vSwitch, where traffic is forwarded directly
481 between VF interfaces by packet forwarder (e.g. testpmd application)
482 * tests without vSwitch, where VM accesses VF interfaces directly
483 by PCI-passthrough_ to measure raw VM throughput performance.
487 Using QEMU with PCI passthrough support
488 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
490 Raw virtual machine throughput performance can be measured by execution of PVP
491 test with direct access to NICs by PCI pass-through. To execute VM with direct
492 access to PCI devices, enable vfio-pci_. In order to use virtual functions,
493 SRIOV-support_ must be enabled.
495 Execution of test with PCI pass-through with vswitch disabled:
497 .. code-block:: console
499 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
500 --vswitch none --vnf QemuPciPassthrough pvp_tput
502 Any of supported guest-loopback-application_ can be used inside VM with
503 PCI pass-through support.
505 Note: Qemu with PCI pass-through support can be used only with PVP test
508 .. _guest-loopback-application:
510 Selection of loopback application for tests with VMs
511 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
513 To select the loopback applications which will forward packets inside VMs,
514 the following parameter should be configured:
516 .. code-block:: python
518 GUEST_LOOPBACK = ['testpmd']
520 or use ``--test-params`` CLI argument:
522 .. code-block:: console
524 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf \
525 --test-params "GUEST_LOOPBACK=['testpmd']"
527 Supported loopback applications are:
529 .. code-block:: console
531 'testpmd' - testpmd from dpdk will be built and used
532 'l2fwd' - l2fwd module provided by Huawei will be built and used
533 'linux_bridge' - linux bridge will be configured
534 'buildin' - nothing will be configured by vsperf; VM image must
535 ensure traffic forwarding between its interfaces
537 Guest loopback application must be configured, otherwise traffic
538 will not be forwarded by VM and testcases with VM related deployments
539 will fail. Guest loopback application is set to 'testpmd' by default.
541 **NOTE:** In case that only 1 or more than 2 NICs are configured for VM,
542 then 'testpmd' should be used. As it is able to forward traffic between
543 multiple VM NIC pairs.
545 **NOTE:** In case of linux_bridge, all guest NICs are connected to the same
546 bridge inside the guest.
548 Mergable Buffers Options with QEMU
549 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
551 Mergable buffers can be disabled with VSPerf within QEMU. This option can
552 increase performance significantly when not using jumbo frame sized packets.
553 By default VSPerf disables mergable buffers. If you wish to enable it you
554 can modify the setting in the a custom conf file.
556 .. code-block:: python
558 GUEST_NIC_MERGE_BUFFERS_DISABLE = [False]
560 Then execute using the custom conf file.
562 .. code-block:: console
564 $ ./vsperf --conf-file=<path_to_custom_conf>/10_custom.conf
566 Alternatively you can just pass the param during execution.
568 .. code-block:: console
570 $ ./vsperf --test-params "GUEST_NIC_MERGE_BUFFERS_DISABLE=[False]"
573 Selection of dpdk binding driver for tests with VMs
574 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
576 To select dpdk binding driver, which will specify which driver the vm NICs will
577 use for dpdk bind, the following configuration parameter should be configured:
579 .. code-block:: console
581 GUEST_DPDK_BIND_DRIVER = ['igb_uio_from_src']
583 The supported dpdk guest bind drivers are:
585 .. code-block:: console
587 'uio_pci_generic' - Use uio_pci_generic driver
588 'igb_uio_from_src' - Build and use the igb_uio driver from the dpdk src
590 'vfio_no_iommu' - Use vfio with no iommu option. This requires custom
591 guest images that support this option. The default
592 vloop image does not support this driver.
594 Note: uio_pci_generic does not support sr-iov testcases with guests attached.
595 This is because uio_pci_generic only supports legacy interrupts. In case
596 uio_pci_generic is selected with the vnf as QemuPciPassthrough it will be
597 modified to use igb_uio_from_src instead.
599 Note: vfio_no_iommu requires kernels equal to or greater than 4.5 and dpdk
600 16.04 or greater. Using this option will also taint the kernel.
602 Please refer to the dpdk documents at http://dpdk.org/doc/guides for more
603 information on these drivers.
605 Guest Core and Thread Binding
606 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
608 VSPERF provides options to achieve better performance by guest core binding and
609 guest vCPU thread binding as well. Core binding is to bind all the qemu threads.
610 Thread binding is to bind the house keeping threads to some CPU and vCPU thread to
611 some other CPU, this helps to reduce the noise from qemu house keeping threads.
614 .. code-block:: python
616 GUEST_CORE_BINDING = [('#EVAL(6+2*#VMINDEX)', '#EVAL(7+2*#VMINDEX)')]
618 **NOTE** By default the GUEST_THREAD_BINDING will be none, which means same as
619 the GUEST_CORE_BINDING, i.e. the vcpu threads are sharing the physical CPUs with
620 the house keeping threads. Better performance using vCPU thread binding can be
621 achieved by enabling affinity in the custom configuration file.
623 For example, if an environment requires 32,33 to be core binded and 29,30&31 for
624 guest thread binding to achieve better performance.
626 .. code-block:: python
628 VNF_AFFINITIZATION_ON = True
629 GUEST_CORE_BINDING = [('32','33')]
630 GUEST_THREAD_BINDING = [('29', '30', '31')]
635 QEMU default to a compatible subset of performance enhancing cpu features.
636 To pass all available host processor features to the guest.
638 .. code-block:: python
640 GUEST_CPU_OPTIONS = ['host,migratable=off']
642 **NOTE** To enhance the performance, cpu features tsc deadline timer for guest,
643 the guest PMU, the invariant TSC can be provided in the custom configuration file.
645 Multi-Queue Configuration
646 ^^^^^^^^^^^^^^^^^^^^^^^^^
648 VSPerf currently supports multi-queue with the following limitations:
650 1. Requires QEMU 2.5 or greater and any OVS version higher than 2.5. The
651 default upstream package versions installed by VSPerf satisfies this
654 2. Guest image must have ethtool utility installed if using l2fwd or linux
655 bridge inside guest for loopback.
657 3. If using OVS versions 2.5.0 or less enable old style multi-queue as shown
658 in the ''02_vswitch.conf'' file.
660 .. code-block:: python
662 OVS_OLD_STYLE_MQ = True
664 To enable multi-queue for dpdk modify the ''02_vswitch.conf'' file.
666 .. code-block:: python
668 VSWITCH_DPDK_MULTI_QUEUES = 2
670 **NOTE:** you should consider using the switch affinity to set a pmd cpu mask
671 that can optimize your performance. Consider the numa of the NIC in use if this
672 applies by checking /sys/class/net/<eth_name>/device/numa_node and setting an
673 appropriate mask to create PMD threads on the same numa node.
675 When multi-queue is enabled, each dpdk or dpdkvhostuser port that is created
676 on the switch will set the option for multiple queues. If old style multi queue
677 has been enabled a global option for multi queue will be used instead of the
680 To enable multi-queue on the guest modify the ''04_vnf.conf'' file.
682 .. code-block:: python
684 GUEST_NIC_QUEUES = [2]
686 Enabling multi-queue at the guest will add multiple queues to each NIC port when
687 qemu launches the guest.
689 In case of Vanilla OVS, multi-queue is enabled on the tuntap ports and nic
690 queues will be enabled inside the guest with ethtool. Simply enabling the
691 multi-queue on the guest is sufficient for Vanilla OVS multi-queue.
693 Testpmd should be configured to take advantage of multi-queue on the guest if
694 using DPDKVhostUser. This can be done by modifying the ''04_vnf.conf'' file.
696 .. code-block:: python
698 GUEST_TESTPMD_PARAMS = ['-l 0,1,2,3,4 -n 4 --socket-mem 512 -- '
699 '--burst=64 -i --txqflags=0xf00 '
700 '--nb-cores=4 --rxq=2 --txq=2 '
703 **NOTE:** The guest SMP cores must be configured to allow for testpmd to use the
704 optimal number of cores to take advantage of the multiple guest queues.
706 In case of using Vanilla OVS and qemu virtio-net you can increase performance
707 by binding vhost-net threads to cpus. This can be done by enabling the affinity
708 in the ''04_vnf.conf'' file. This can be done to non multi-queue enabled
709 configurations as well as there will be 2 vhost-net threads.
711 .. code-block:: python
713 VSWITCH_VHOST_NET_AFFINITIZATION = True
715 VSWITCH_VHOST_CPU_MAP = [4,5,8,11]
717 **NOTE:** This method of binding would require a custom script in a real
720 **NOTE:** For optimal performance guest SMPs and/or vhost-net threads should be
721 on the same numa as the NIC in use if possible/applicable. Testpmd should be
722 assigned at least (nb_cores +1) total cores with the cpu mask.
727 VSPERF provides options to support jumbo frame testing with a jumbo frame supported
728 NIC and traffic generator for the following vswitches:
734 3. TestPMD loopback with or without a guest
736 **NOTE:** There is currently no support for SR-IOV or VPP at this time with jumbo
739 All packet forwarding applications for pxp testing is supported.
741 To enable jumbo frame testing simply enable the option in the conf files and set the
742 maximum size that will be used.
744 .. code-block:: python
746 VSWITCH_JUMBO_FRAMES_ENABLED = True
747 VSWITCH_JUMBO_FRAMES_SIZE = 9000
749 To enable jumbo frame testing with OVSVanilla the NIC in test on the host must have
750 its mtu size changed manually using ifconfig or applicable tools:
752 .. code-block:: console
754 ifconfig eth1 mtu 9000 up
756 **NOTE:** To make the setting consistent across reboots you should reference the OS
757 documents as it differs from distribution to distribution.
759 To start a test for jumbo frames modify the conf file packet sizes or pass the option
760 through the VSPERF command line.
762 .. code-block:: python
764 TEST_PARAMS = {'TRAFFICGEN_PKT_SIZES':(2000,9000)}
766 .. code-block:: python
768 ./vsperf --test-params "TRAFFICGEN_PKT_SIZES=2000,9000"
770 It is recommended to increase the memory size for OvsDpdkVhostUser testing from the default
771 1024. Your size required may vary depending on the number of guests in your testing. 4096
772 appears to work well for most typical testing scenarios.
774 .. code-block:: python
776 DPDK_SOCKET_MEM = ['4096', '0']
778 **NOTE:** For Jumbo frames to work with DpdkVhostUser, mergable buffers will be enabled by
779 default. If testing with mergable buffers in QEMU is desired, disable Jumbo Frames and only
780 test non jumbo frame sizes. Test Jumbo Frames sizes separately to avoid this collision.
783 Executing Packet Forwarding tests
784 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
786 To select the applications which will forward packets,
787 the following parameters should be configured:
789 .. code-block:: python
794 or use ``--vswitch`` and ``--fwdapp`` CLI arguments:
796 .. code-block:: console
798 $ ./vsperf phy2phy_cont --conf-file user_settings.py \
802 Supported Packet Forwarding applications are:
804 .. code-block:: console
806 'testpmd' - testpmd from dpdk
809 1. Update your ''10_custom.conf'' file to use the appropriate variables
810 for selected Packet Forwarder:
812 .. code-block:: python
814 # testpmd configuration
816 # packet forwarding mode supported by testpmd; Please see DPDK documentation
817 # for comprehensive list of modes supported by your version.
818 # e.g. io|mac|mac_retry|macswap|flowgen|rxonly|txonly|csum|icmpecho|...
819 # Note: Option "mac_retry" has been changed to "mac retry" since DPDK v16.07
820 TESTPMD_FWD_MODE = 'csum'
821 # checksum calculation layer: ip|udp|tcp|sctp|outer-ip
822 TESTPMD_CSUM_LAYER = 'ip'
823 # checksum calculation place: hw (hardware) | sw (software)
824 TESTPMD_CSUM_CALC = 'sw'
825 # recognize tunnel headers: on|off
826 TESTPMD_CSUM_PARSE_TUNNEL = 'off'
830 .. code-block:: console
832 $ ./vsperf phy2phy_tput --conf-file <path_to_settings_py>
834 Executing Packet Forwarding tests with one guest
835 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
837 TestPMD with DPDK 16.11 or greater can be used to forward packets as a switch to a single guest using TestPMD vdev
838 option. To set this configuration the following parameters should be used.
840 .. code-block:: python
845 or use ``--vswitch`` and ``--fwdapp`` CLI arguments:
847 .. code-block:: console
849 $ ./vsperf pvp_tput --conf-file user_settings.py \
853 Guest forwarding application only supports TestPMD in this configuration.
855 .. code-block:: python
857 GUEST_LOOPBACK = ['testpmd']
859 For optimal performance one cpu per port +1 should be used for TestPMD. Also set additional params for packet forwarding
860 application to use the correct number of nb-cores.
862 .. code-block:: python
864 DPDK_SOCKET_MEM = ['1024', '0']
865 VSWITCHD_DPDK_ARGS = ['-l', '46,44,42,40,38', '-n', '4']
866 TESTPMD_ARGS = ['--nb-cores=4', '--txq=1', '--rxq=1']
868 For guest TestPMD 3 VCpus should be assigned with the following TestPMD params.
870 .. code-block:: python
872 GUEST_TESTPMD_PARAMS = ['-l 0,1,2 -n 4 --socket-mem 1024 -- '
873 '--burst=64 -i --txqflags=0xf00 '
874 '--disable-hw-vlan --nb-cores=2 --txq=1 --rxq=1']
876 Execution of TestPMD can be run with the following command line
878 .. code-block:: console
880 ./vsperf pvp_tput --vswitch=none --fwdapp=TestPMD --conf-file <path_to_settings_py>
882 **NOTE:** To achieve the best 0% loss numbers with rfc2544 throughput testing, other tunings should be applied to host
883 and guest such as tuned profiles and CPU tunings to prevent possible interrupts to worker threads.
885 VSPERF modes of operation
886 ^^^^^^^^^^^^^^^^^^^^^^^^^
888 VSPERF can be run in different modes. By default it will configure vSwitch,
889 traffic generator and VNF. However it can be used just for configuration
890 and execution of traffic generator. Another option is execution of all
891 components except traffic generator itself.
893 Mode of operation is driven by configuration parameter -m or --mode
895 .. code-block:: console
897 -m MODE, --mode MODE vsperf mode of operation;
899 "normal" - execute vSwitch, VNF and traffic generator
900 "trafficgen" - execute only traffic generator
901 "trafficgen-off" - execute vSwitch and VNF
902 "trafficgen-pause" - execute vSwitch and VNF but wait before traffic transmission
904 In case, that VSPERF is executed in "trafficgen" mode, then configuration
905 of traffic generator can be modified through ``TRAFFIC`` dictionary passed to the
906 ``--test-params`` option. It is not needed to specify all values of ``TRAFFIC``
907 dictionary. It is sufficient to specify only values, which should be changed.
908 Detailed description of ``TRAFFIC`` dictionary can be found at
909 :ref:`configuration-of-traffic-dictionary`.
911 Example of execution of VSPERF in "trafficgen" mode:
913 .. code-block:: console
915 $ ./vsperf -m trafficgen --trafficgen IxNet --conf-file vsperf.conf \
916 --test-params "TRAFFIC={'traffic_type':'rfc2544_continuous','bidir':'False','framerate':60}"
918 Code change verification by pylint
919 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
921 Every developer participating in VSPERF project should run
922 pylint before his python code is submitted for review. Project
923 specific configuration for pylint is available at 'pylint.rc'.
925 Example of manual pylint invocation:
927 .. code-block:: console
929 $ pylint --rcfile ./pylintrc ./vsperf
934 Custom image fails to boot
935 ~~~~~~~~~~~~~~~~~~~~~~~~~~
937 Using custom VM images may not boot within VSPerf pxp testing because of
938 the drive boot and shared type which could be caused by a missing scsi
939 driver inside the image. In case of issues you can try changing the drive
942 .. code-block:: python
944 GUEST_BOOT_DRIVE_TYPE = ['ide']
945 GUEST_SHARED_DRIVE_TYPE = ['ide']
947 OVS with DPDK and QEMU
948 ~~~~~~~~~~~~~~~~~~~~~~~
950 If you encounter the following error: "before (last 100 chars):
951 '-path=/dev/hugepages,share=on: unable to map backing store for
952 hugepages: Cannot allocate memory\r\n\r\n" during qemu initialization,
953 check the amount of hugepages on your system:
955 .. code-block:: console
957 $ cat /proc/meminfo | grep HugePages
960 By default the vswitchd is launched with 1Gb of memory, to change
961 this, modify --socket-mem parameter in conf/02_vswitch.conf to allocate
962 an appropriate amount of memory:
964 .. code-block:: python
966 DPDK_SOCKET_MEM = ['1024', '0']
967 VSWITCHD_DPDK_ARGS = ['-c', '0x4', '-n', '4']
968 VSWITCHD_DPDK_CONFIG = {
969 'dpdk-init' : 'true',
970 'dpdk-lcore-mask' : '0x4',
971 'dpdk-socket-mem' : '1024,0',
974 Note: Option ``VSWITCHD_DPDK_ARGS`` is used for vswitchd, which supports ``--dpdk``
975 parameter. In recent vswitchd versions, option ``VSWITCHD_DPDK_CONFIG`` will be
976 used to configure vswitchd via ``ovs-vsctl`` calls.
982 For more information and details refer to the rest of vSwitchPerfuser documentation.