Age | Commit message (Collapse) | Author | Files | Lines |
|
Changes to IxNextgen API:
- modify API for traffic model configuration so that it can be used
to create/configure different models
- add methods to start and stop protocols on IxNetwork
- add method for getting assigned ports
JIRA: YARDSTICK-1455
Change-Id: I43c10fea9ba83994caf373ceff8b0252a45439b3
Signed-off-by: Serhiy Pshyk <serhiyx.pshyk@intel.com>
(cherry picked from commit 7237241a396871d637f9906c1fc93a8587a7b726)
|
|
|
|
|
|
Using AgnosticVNF is now possible to create a "black box"
testcase for any VNF.
Example rfc2544 test with TRex as trafficgen
samples/vnf_samples/nsut/agnostic/tc_baremetal_rfc2544_ipv4_64B_trex.yaml
and Ixia
samples/vnf_samples/nsut/agnostic/tc_baremetal_rfc2544_latency_ipv4_64B_ixia.yaml
NOTE:
Yardstick will not interact with VNF.
No KPIs from VNF will be collected
JIRA: YARDSTICK-1438
Change-Id: I408b3dcb90f29c7b41fcf8f883b5abfc4e9c3dbc
Signed-off-by: Chornyi, TarasX <tarasx.chornyi@intel.com>
|
|
'YARDSTICK-1359', 'YARDSTICK-1356'
* changes:
Add vEPC service request test cases
Add network initiated dedicated bearers creation
Add UE initiated dedicated bearer creation test
Add vEPC infrastructure for Landslide TG
Add vEPC default bearers relocation test case
Add vEPC default bearers create/delete test case
Add Spirent Landslide traffic profile templates
Update LSResourceHelper unittests
Add Spirent Landslide TG API
Added Landslide Resource Helper implementation
Add TclClients for Spirent Landslide TG
|
|
In current implementation of Yardstick, SA context uses the
same SSH key file name for each VM created. E.g. the Yardstick
overrides the key file context each time the new VM is created.
As a result, the key file is valid only for last VM created
by Yardstick causing the connection problem for other VMs.
Solution is to generate different key file name for each VM
name separately.
JIRA: YARDSTICK-1436
Change-Id: I22ae2874695736255aa3314654a2cced9338b53d
Signed-off-by: Mytnyk, Volodymyr <volodymyrx.mytnyk@intel.com>
|
|
New IxNextgen API allows to configure Vlans(vlan id, priority, tp id)
to Ethernet layer of device group.
JIRA: YARDSTICK-1420
Change-Id: I1c5fe37b5beb11daf1c539933fcad60832553bb9
Signed-off-by: Serhiy Pshyk <serhiyx.pshyk@intel.com>
|
|
Add UE and network initiated testcase and related session profiles.
JIRA: YARDSTICK-1429
Change-Id: I16d6c065cf346db31e235ac5add57509cfbc8009
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
Adding related pod file, session profile and testcase file.
JIRA: YARDSTICK-1355
Change-Id: Ic97d59ef3a36ac05c6c369c36f917ae95d18d863
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
Adding session profile and test case file.
JIRA: YARDSTICK-1353
Change-Id: Iaa98fd8cb3ed243f9ac157089de32fa23ec0e6b4
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
1. Add yardstick topology and model for fully emulated environment.
2. Add support for black-box testing of real vEPC VNF vs Spirent Landslide
traffic generator:
- vEPC VNF pod file
- topology and model files aware of vEPC VNF connection to Spirent
Landslide TG
- stub class for vEPC VNF, and related unit tests
JIRA: YARDSTICK-1424
Change-Id: If7b6d19919679347c4360f4a0f2e420716e7fd2d
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
This test case runs on Spirent Landslide TG. All vEPC entities are
emulated on Landslide test servers.
Test case actions sequence:
- create default bearers
- start sending traffic
- partial/full bearers relocation between two ENodeB's, Intra-MME
- disconnect all bearers
JIRA: YARDSTICK-1354
Change-Id: I95c7b0001ddb363402013d83ae8617b88b2a4bd9
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
This test case runs on Spirent Landslide TG. All vEPC entities are
emulated on Landslide test server.
Test run configuration is performed in following files:
1. test case file
2. pod file
3. session profile file (landslide_session_default_bearer*.yaml)
The structure of these files represents configuration of different
parts of test cases in this test session. The order of test cases should
always match.
Spirent Landslide test types involved: "SGW_Node", "SGW_Nodal".
Test case actions sequence:
- create default bearers (as per value of "Sessions" key) with
specified rate (session profile key "StartRate")
- send traffic flow(s) (pre-configured in traffic profile) for a number
of seconds (specified in session profile key "duration")
- remove N default bearers with specified rate (session profile key
"DisconnectRate")
Parameters with ALL_CAPS values are placeholders.
User should replace them with their own condifuration.
Session profile file placeholders do not need to be replaced as long as
these parameters are overriden in other files (e.g. pod, test case).
Test results are dispatched according to runner configuration options.
For fully emulated test runs like this one, best suited runner is "Search"
type. In case of testing against real EPC VNFs, user could prefer
different runner type.
JIRA: YARDSTICK-1348
Change-Id: I5caf4878fb212b38d4148a297c5dba03b8ab98a8
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
These files are templates of Spirent Landslide data message flow (DMF)
representing traffic profile of one or multiple traffic flows.
Template files should not be edited directly. Instead, their parameters
could be overriden in test case YAML file, in section "dmf".
JIRA: YARDSTICK-1359
Change-Id: I4980a784a78ca870c18457227248e6e2ac0b8eda
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
Soften sed pattern, in order to cope with new contents of
ansible/install.yaml, recently modified by change 61173.
Previous sed pattern does not work anymore against this changed line:
https://gerrit.opnfv.org/gerrit/#/c/61173/4/ansible/install.yaml@39
JIRA: YARDSTICK-1411
Change-Id: I1e1ebd454900252c29263e746470e18e3139e016
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
|
|
Assigning IXIA ports one by one takes more time than
assign all ports by once. Also, assigning ports in
one call resolves the assign port issues caused on
some IXIA chassis versions.
JIRA: YARDSTICK-1425
Change-Id: Ibb08e40ee1a8762573f6752bdd1f583c5a1d9d2a
Signed-off-by: Mytnyk, Volodymyr <volodymyrx.mytnyk@intel.com>
Signed-off-by: Rodolfo Alonso Hernandez <rodolfo.alonso.hernandez@intel.com>
|
|
New unit tests have been added to make sure the right
commands are sent to Spirent LS.
This helps to ensure that the correct commands are sent
to Spirent, so that potentially breaking changes can be
flagged.
JIRA: YARDSTICK-1356
Change-Id: I23f452355afcec2705da7d4665a34b178e2ba729
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
|
|
IXIA API sets the src/dst mac of the traffic always into
default value. The `srcmac`/`dstmac` is never set in `traffic_param`
map [1]. Instead the src/dst mac is stored in traffic_param[`outer_l2`].
[1] https://gerrit.opnfv.org/gerrit/gitweb?p=yardstick.git;a=blob;f=yardstick/network_services/traffic_profile/ixia_rfc2544.py;h=44bf2eafc7f8cf03ebd2c8520d316358fe1e5fc2;hb=HEAD#l92
JIRA: YARDSTICK-1423
Change-Id: Iabab7a54404b284d3b95bbea6c8624233c138351
Signed-off-by: Mytnyk, Volodymyr <volodymyrx.mytnyk@intel.com>
|
|
|
|
and requirements.txt""
|
|
|
|
|
|
|
|
New TG class "LandslideTrafficGen" represents an interface to use Spirent
Landslide API to configure and execute vEPC test cases on
hardware/software Spirent environment. For that purpose this class and its
helper classes use Spirent Landslide REST API calls and Landslide TCL API
calls via library module lsapi.py.
Change-Id: Ib6560d5cd2483c6c9f5c95568345ac39bfebbd4d
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
Now "daemonize" prevents from installing Yardstick on baremetal.
Configuring:
"daemonize" in case "container" installation mode;
"logto" in case "baremetal" installation mode;
JIRA: YARDSTICK-1395
Change-Id: Ie390464db9c0816583edc711ed6526629d091bd3
Signed-off-by: Stepan Andrushko <stepanx.andrushko@intel.com>
|
|
|
|
make samples/fio.yaml to support both Openstack and K8s
with context_type specified. It defaultly use Heat context, if
for k8s,
yardstick -d task start samples/fio.yaml --task-args
'{"context_type": "Kubernetes"}'
JIRA: YARDSTICK-1309
Change-Id: I2b0d61cf2f197ec8c60bda0ab4ff51594709f482
Signed-off-by: rexlee8776 <limingjiang@huawei.com>
|
|
|
|
make samples/lmbench.yaml to support both Openstack and K8s
with context_type specified. It defaultly use Heat context, if
for k8s,
yardstick -d task start samples/lmbench.yaml --task-args
'{"context_type": "Kubernetes"}'
JIRA: YARDSTICK-1310
Change-Id: Ifa542e830fb5ab7823bd4f994ab5d7a8e8a5b415
Signed-off-by: rexlee8776 <limingjiang@huawei.com>
|
|
|
|
requirements.txt"
This reverts commit 662de49b9f6c116d15882aed9e678cd3cf67494b.
JIRA: YARDSTICK-1411
Change-Id: Id277aa8711c9ff77e085d26c90546eecc3490535
Signed-off-by: rexlee8776 <limingjiang@huawei.com>
|
|
Change-Id: If01abfd5a99b4cea74f835e258f3a982af9446cd
Signed-off-by: rexlee8776 <limingjiang@huawei.com>
|
|
Class "LandslideResourceHelper" provides API for operations needed
- to configure Landslide test session
- manage test session execution (start/stop/abort)
- collect measurements during test run
This helper class API is responsible to for configure Landslide test runs:
- create test user
- create test servers (for emulation of specific vEPC blocks)
- create SUTs (actual tested VNFs performing specific vEPC roles)
- create test session (contains actual test cases)
- create DMFs (pre-defined traffic flows in traffic profile)
- operate traffic run execution (start, stop, abort)
- monitor test run status
- collect KPIs on TG side
Some of these features use Landslide REST API.
Other ones use Landslide TCL API.
JIRA: YARDSTICK-1356
Change-Id: I8fc8a7d85301121da465d054b8d38ae09a541c36
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
New IxNextgen API allows to perform the following configurations on IxNetwork:
- add topology
- add device group to topology
- add ethernet protocol layer to device group
- add ipv4 protocol layer to device group
- add pppox protocol layer to device group
JIRA: YARDSTICK-1408
Change-Id: Icf5bb5f24784cf2c167c515b04c81471ac425e38
Signed-off-by: Serhiy Pshyk <serhiyx.pshyk@intel.com>
|
|
JIRA: YARDSTICK-1410
Change-Id: I23b4c0bcfa9aa431e259930a41d4171ff5b4e8a1
Signed-off-by: rexlee8776 <limingjiang@huawei.com>
|
|
Add class "LandslideTclClient" which holds TCL calls wrappers for
operations on the LandSlide TG.
It also introduces several new parameters in
- test case YAML file structure:
* session_profile: YAML file holding a template of Landslide test
session for specific vEPC test scenario
* dmf: test case options member holding traffic flow specific parameters
for designated test scenario, that override ones in traffic_profile
YAML file
* test_cases: test case options member holding parameters specific to
designated vEPC test case, overriding those in session_profile
template file
- pod YAML file structure:
* tas_manager: holds parameters used to connect to Spirent Landslide TAS
* config: list of entries describing Landslide test servers access,
their network interfaces configuration, information of vEPC elements
emulated by specific test server
Class "LandslideProfile" implements access to traffic profile template
from "LandslideResourceHelper" class.
JIRA: YARDSTICK-1356
Change-Id: Ia6428072b1e72025be17c94e5141bd23f7993247
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
Signed-off-by: Orest Voznyy <orestx.voznyy@intel.com>
|
|
CLI argument $1 was used for both admin-openrc file with OpenStack
credentials and Yardstick Docker image, so obviously one of them would
be wrong, and NSB setup would fail at some point.
Historically, admin-openrc could be supplied as CLI argument $1, but
change https://gerrit.opnfv.org/gerrit/56493 broke it, which means
specifying Yardstick Docker image as CLI argument never worked.
This fix implements support for the following CLI options:
-o openrc Specify admin-openrc file with OpenStack credentials
-i image Specify Yardstick Docker image
-h Provide usage information
For backward compatibility reasons, admin-openrc file can be supplied
either with the -o option, or as a CLI argument.
JIRA: YARDSTICK-1137
Change-Id: I3e9904970560468c24b21738203bc67b7591f9e4
Signed-off-by: Patrice Buriez <patrice.buriez@intel.com>
|
|
|
|
|
|
Come changes re: Status == "Result" and Status == "Success" and
L3FWD changes onto latest L3FWD Dashboard
Change-Id: I62d0cfc30167085ad5efe348ab118f1defbc4901
Signed-off-by: Daniel Martin Buckley <daniel.m.buckley@intel.com>
|
|
JIRA: YARDSTICK-1404
The test scripts require an extra argument to support collectd.
Also the CPU utilization is a different variable name structure
in the JSON file.
Change-Id: I127b7e86c0f93f7743331a2d71eb81d690e42471
Signed-off-by: Daniel Martin Buckley <daniel.m.buckley@intel.com>
|
|
* Use time.sleep to avoid multiprocessing.Queue broken pipe error.
JIRA: YARDSTICK-1245
Change-Id: I1ef1f8e68a553278996b38c226314c13dd974faa
Signed-off-by: Miikka Koistinen <miikka.koistinen@nokia.com>
|
|
Function "_create_imix_data" must return the weight in percentage, not in
absolute value.
JIRA: YARDSTICK-1406
Change-Id: I3bc409bab17ae5778a6783f4319699009a5437fa
Signed-off-by: Rodolfo Alonso Hernandez <rodolfo.alonso.hernandez@intel.com>
|
|
Supported Yardstick installation modes in yardstick.yaml: baremetal
and container. First is used to install Yardstick on baremetal,
another is used during docker build to install Yardstick in container.
Added third option: 'container pull': to pull docker image
(opnfv/yardstick) from Docker hub and start container.
JIRA: YARDSTICK-1392
Change-Id: I4158de3eb6fcbe0018a15fdda89a3e3ab29c26bb
Signed-off-by: Stepan Andrushko <stepanx.andrushko@intel.com>
|
|
|
|
|
|
Add possibility ro reuse infra_deploy.yaml to spawn VMs on jump host,
depending on jump host Ubuntu version: either 16 or 18.
JIRA: YARDSTICK-1396
Change-Id: Id0682c03098e611fe3a51cbd24d8ff2dda7c552a
Signed-off-by: Stepan Andrushko <stepanx.andrushko@intel.com>
|
|
|