diff options
35 files changed, 945 insertions, 904 deletions
@@ -4,6 +4,7 @@ .tox .cache .eggs +.pytest_cache/ venv nfvbench.egg-info nfvbenchvm/dib/dib-venv @@ -1,14 +1,27 @@ NFVbench: A Network Performance Benchmarking Tool for NFVi Full Stacks ********************************************************************** -The NFVbench tool provides an automated way to measure the network performance for the most common data plane packet flows on any OpenStack based NFVi system viewed as a black box (NFVi Full Stack). +The NFVbench tool provides an automated way to measure the network performance for the most common data plane packet flows +on any NFVi system viewed as a black box (NFVi Full Stack). An NFVi full stack exposes the following interfaces: -- an OpenStack API +- an OpenStack API for those NFVi platforms based on OpenStack - an interface to send and receive packets on the data plane (typically through top of rack switches) -The NFVi full stack does not have to be supported by the OPNFV ecosystem and can be any functional OpenStack system that provides the aboce interfaces. NFVbench can be installed standalone (in the form of a single Docker container) and is fully functional without the need to install any other OPNFV framework. +The NFVi full stack does not have to be supported by the OPNFV ecosystem and can be any functional OpenStack system that provides +the above interfaces. +NFVbench can also be used without OpenStack on any bare system that can handle L2 forwarding or L3 routing. -It is designed to be easy to install and easy to use by non experts (no need to be an expert in traffic generators and data plane performance testing). +NFVbench can be installed standalone (in the form of a single Docker container) and is fully functional without +the need to install any other OPNFV framework. + +It is designed to be easy to install and easy to use by non experts (no need to be an expert in traffic generators and data plane +performance benchmarking). + +Online Documentation +-------------------- +The latest version of the NFVbench documentation is available online at: + +http://docs.opnfv.org/en/latest/submodules/nfvbench/docs/testing/user/userguide/index.html Contact Information @@ -16,4 +29,3 @@ Contact Information Inquiries and questions: send an email to opnfv-tech-discuss@lists.opnfv.org with a Subject line starting with "[nfvbench]". Open issues or submit an issue or enhancement request: https://jira.opnfv.org/projects/NFVBENCH/issues (this requires an OPNFV Linux Foundation login). - diff --git a/cleanup/__init__.py b/cleanup/__init__.py deleted file mode 100644 index 04924ec..0000000 --- a/cleanup/__init__.py +++ /dev/null @@ -1,13 +0,0 @@ -# Copyright 2017 Cisco Systems, Inc. All rights reserved. -# -# Licensed under the Apache License, Version 2.0 (the "License"); you may -# not use this file except in compliance with the License. You may obtain -# a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT -# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the -# License for the specific language governing permissions and limitations -# under the License. diff --git a/cleanup/nfvbench_cleanup.py b/cleanup/nfvbench_cleanup.py deleted file mode 100644 index 1520647..0000000 --- a/cleanup/nfvbench_cleanup.py +++ /dev/null @@ -1,594 +0,0 @@ -#!/usr/bin/env python -# Copyright 2017 Cisco Systems, Inc. All rights reserved. -# -# Licensed under the Apache License, Version 2.0 (the "License"); you may -# not use this file except in compliance with the License. You may obtain -# a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT -# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the -# License for the specific language governing permissions and limitations -# under the License. -# - -############################################################################### -# # -# This is a helper script which will delete all resources created by # -# NFVbench. # -# # -# Normally, NFVbench will clean up automatically when it is done. However, # -# sometimes errors or timeouts happen during the resource creation stage, # -# which will cause NFVbench out of sync with the real environment. If that # -# happens, a force cleanup may be needed. # -# # -# It is safe to use the script with the resource list generated by # -# NFVbench, usage: # -# $ python nfvbench_cleanup.py -r /path/to/openrc # -# # -# Note: If running under single-tenant or tenant/user reusing mode, you have # -# to cleanup the server resources first, then client resources. # -# # -# When there is no resource list provided, the script will simply grep the # -# resource name with "nfvbench" and delete them. If running on a production # -# network, please double and triple check all resources names are *NOT* # -# starting with "nfvbench", otherwise they will be deleted by the script. # -# # -############################################################################### - -# ====================================================== -# WARNING -# ====================================================== -# IMPORTANT FOR RUNNING NFVbench ON PRODUCTION CLOUDS -# -# DOUBLE CHECK THE NAMES OF ALL RESOURCES THAT DO NOT -# BELONG TO NFVbench ARE *NOT* STARTING WITH "nfvbench". -# ====================================================== - -from abc import ABCMeta -from abc import abstractmethod -import argparse -import re -import sys -import time -import traceback - -# openstack python clients -import cinderclient -from keystoneclient import client as keystoneclient -import neutronclient -from novaclient.exceptions import NotFound -from tabulate import tabulate - -from nfvbench import credentials - -resource_name_re = None - -def prompt_to_run(): - print "Warning: You didn't specify a resource list file as the input. "\ - "The script will delete all resources shown above." - answer = raw_input("Are you sure? (y/n) ") - if answer.lower() != 'y': - sys.exit(0) - -def fetch_resources(fetcher, options=None): - try: - if options: - res_list = fetcher(search_opts=options) - else: - res_list = fetcher() - except Exception as e: - res_list = [] - traceback.print_exc() - print "Warning exception while listing resources:" + str(e) - resources = {} - for res in res_list: - # some objects provide direct access some - # require access by key - try: - resid = res.id - resname = res.name - except AttributeError: - resid = res['id'] - resname = res['name'] - if resname and resource_name_re.match(resname): - resources[resid] = resname - return resources - -class AbstractCleaner(object): - __metaclass__ = ABCMeta - - def __init__(self, res_category, res_desc, resources, dryrun): - self.dryrun = dryrun - self.category = res_category - self.resources = {} - if not resources: - print 'Discovering %s resources...' % (res_category) - for rtype, fetch_args in res_desc.iteritems(): - if resources: - if rtype in resources: - self.resources[rtype] = resources[rtype] - else: - self.resources[rtype] = fetch_resources(*fetch_args) - - def report_deletion(self, rtype, name): - if self.dryrun: - print ' + ' + rtype + ' ' + name + ' should be deleted (but is not deleted: dry run)' - else: - print ' + ' + rtype + ' ' + name + ' is successfully deleted' - - def report_not_found(self, rtype, name): - print ' ? ' + rtype + ' ' + name + ' not found (already deleted?)' - - def report_error(self, rtype, name, reason): - print ' - ' + rtype + ' ' + name + ' ERROR:' + reason - - def get_resource_list(self): - result = [] - for rtype, rdict in self.resources.iteritems(): - for resid, resname in rdict.iteritems(): - result.append([rtype, resname, resid]) - return result - - @abstractmethod - def clean(self): - pass - -class StorageCleaner(AbstractCleaner): - def __init__(self, sess, resources, dryrun): - from cinderclient import client as cclient - from novaclient import client as nclient - - self.nova = nclient.Client('2', endpoint_type='publicURL', session=sess) - self.cinder = cclient.Client('2', endpoint_type='publicURL', session=sess) - - res_desc = {'volumes': [self.cinder.volumes.list, {"all_tenants": 1}]} - super(StorageCleaner, self).__init__('Storage', res_desc, resources, dryrun) - - def clean(self): - print '*** STORAGE cleanup' - try: - volumes = [] - detaching_volumes = [] - for id, name in self.resources['volumes'].iteritems(): - try: - vol = self.cinder.volumes.get(id) - if vol.attachments: - # detach the volume - try: - if not self.dryrun: - ins_id = vol.attachments[0]['server_id'] - self.nova.volumes.delete_server_volume(ins_id, id) - print ' . VOLUME ' + vol.name + ' detaching...' - else: - print ' . VOLUME ' + vol.name + ' to be detached...' - detaching_volumes.append(vol) - except NotFound: - print 'WARNING: Volume %s attached to an instance that no longer '\ - 'exists (will require manual cleanup of the database)' % (id) - except Exception as e: - print str(e) - else: - # no attachments - volumes.append(vol) - except cinderclient.exceptions.NotFound: - self.report_not_found('VOLUME', name) - - # check that the volumes are no longer attached - if detaching_volumes: - if not self.dryrun: - print ' . Waiting for %d volumes to be fully detached...' % \ - (len(detaching_volumes)) - retry_count = 5 + len(detaching_volumes) - while True: - retry_count -= 1 - for vol in list(detaching_volumes): - if not self.dryrun: - latest_vol = self.cinder.volumes.get(detaching_volumes[0].id) - if self.dryrun or not latest_vol.attachments: - if not self.dryrun: - print ' + VOLUME ' + vol.name + ' detach complete' - detaching_volumes.remove(vol) - volumes.append(vol) - if detaching_volumes and not self.dryrun: - if retry_count: - print ' . VOLUME %d left to be detached, retries left=%d...' % \ - (len(detaching_volumes), retry_count) - time.sleep(2) - else: - print ' - VOLUME detach timeout, %d volumes left:' % \ - (len(detaching_volumes)) - for vol in detaching_volumes: - print ' ', vol.name, vol.status, vol.id, vol.attachments - break - else: - break - - # finally delete the volumes - for vol in volumes: - if not self.dryrun: - try: - vol.force_delete() - except cinderclient.exceptions.BadRequest as exc: - print str(exc) - self.report_deletion('VOLUME', vol.name) - except KeyError: - pass - -class ComputeCleaner(AbstractCleaner): - def __init__(self, sess, resources, dryrun): - from neutronclient.neutron import client as nclient - from novaclient import client as novaclient - self.neutron_client = nclient.Client('2.0', endpoint_type='publicURL', session=sess) - self.nova_client = novaclient.Client('2', endpoint_type='publicURL', session=sess) - res_desc = { - 'instances': [self.nova_client.servers.list, {"all_tenants": 1}], - 'flavors': [self.nova_client.flavors.list], - 'keypairs': [self.nova_client.keypairs.list] - } - super(ComputeCleaner, self).__init__('Compute', res_desc, resources, dryrun) - - def clean(self): - print '*** COMPUTE cleanup' - try: - # Get a list of floating IPs - fip_lst = self.neutron_client.list_floatingips()['floatingips'] - deleting_instances = self.resources['instances'] - for id, name in self.resources['instances'].iteritems(): - try: - if self.nova_client.servers.get(id).addresses.values(): - ins_addr = self.nova_client.servers.get(id).addresses.values()[0] - fips = [x['addr'] for x in ins_addr if x['OS-EXT-IPS:type'] == 'floating'] - else: - fips = [] - if self.dryrun: - self.nova_client.servers.get(id) - for fip in fips: - self.report_deletion('FLOATING IP', fip) - self.report_deletion('INSTANCE', name) - else: - for fip in fips: - fip_id = [x['id'] for x in fip_lst if x['floating_ip_address'] == fip] - self.neutron_client.delete_floatingip(fip_id[0]) - self.report_deletion('FLOATING IP', fip) - self.nova_client.servers.delete(id) - except NotFound: - deleting_instances.remove(id) - self.report_not_found('INSTANCE', name) - - if not self.dryrun and len(deleting_instances): - print ' . Waiting for %d instances to be fully deleted...' % \ - (len(deleting_instances)) - retry_count = 5 + len(deleting_instances) - while True: - retry_count -= 1 - for ins_id in deleting_instances.keys(): - try: - self.nova_client.servers.get(ins_id) - except NotFound: - self.report_deletion('INSTANCE', deleting_instances[ins_id]) - deleting_instances.pop(ins_id) - - if not len(deleting_instances): - break - - if retry_count: - print ' . INSTANCE %d left to be deleted, retries left=%d...' % \ - (len(deleting_instances), retry_count) - time.sleep(2) - else: - print ' - INSTANCE deletion timeout, %d instances left:' % \ - (len(deleting_instances)) - for ins_id in deleting_instances.keys(): - try: - ins = self.nova_client.servers.get(ins_id) - print ' ', ins.name, ins.status, ins.id - except NotFound: - print(' ', deleting_instances[ins_id], - '(just deleted)', ins_id) - break - except KeyError: - pass - - try: - for id, name in self.resources['flavors'].iteritems(): - try: - flavor = self.nova_client.flavors.find(name=name) - if not self.dryrun: - flavor.delete() - self.report_deletion('FLAVOR', name) - except NotFound: - self.report_not_found('FLAVOR', name) - except KeyError: - pass - - try: - for id, name in self.resources['keypairs'].iteritems(): - try: - if self.dryrun: - self.nova_client.keypairs.get(name) - else: - self.nova_client.keypairs.delete(name) - self.report_deletion('KEY PAIR', name) - except NotFound: - self.report_not_found('KEY PAIR', name) - except KeyError: - pass - -class NetworkCleaner(AbstractCleaner): - - def __init__(self, sess, resources, dryrun): - from neutronclient.neutron import client as nclient - self.neutron = nclient.Client('2.0', endpoint_type='publicURL', session=sess) - - # because the response has an extra level of indirection - # we need to extract it to present the list of network or router objects - def networks_fetcher(): - return self.neutron.list_networks()['networks'] - - def routers_fetcher(): - return self.neutron.list_routers()['routers'] - - def secgroup_fetcher(): - return self.neutron.list_security_groups()['security_groups'] - - res_desc = { - 'sec_groups': [secgroup_fetcher], - 'networks': [networks_fetcher], - 'routers': [routers_fetcher] - } - super(NetworkCleaner, self).__init__('Network', res_desc, resources, dryrun) - - def remove_router_interface(self, router_id, port): - """ - Remove the network interface from router - """ - body = { - # 'port_id': port['id'] - 'subnet_id': port['fixed_ips'][0]['subnet_id'] - } - try: - self.neutron.remove_interface_router(router_id, body) - self.report_deletion('Router Interface', port['fixed_ips'][0]['ip_address']) - except neutronclient.common.exceptions.NotFound: - pass - - def remove_network_ports(self, net): - """ - Remove ports belonging to network - """ - for port in filter(lambda p: p['network_id'] == net, self.neutron.list_ports()['ports']): - try: - self.neutron.delete_port(port['id']) - self.report_deletion('Network port', port['id']) - except neutronclient.common.exceptions.NotFound: - pass - - def clean(self): - print '*** NETWORK cleanup' - - try: - for id, name in self.resources['sec_groups'].iteritems(): - try: - if self.dryrun: - self.neutron.show_security_group(id) - else: - self.neutron.delete_security_group(id) - self.report_deletion('SECURITY GROUP', name) - except NotFound: - self.report_not_found('SECURITY GROUP', name) - except KeyError: - pass - - try: - for id, name in self.resources['floating_ips'].iteritems(): - try: - if self.dryrun: - self.neutron.show_floatingip(id) - else: - self.neutron.delete_floatingip(id) - self.report_deletion('FLOATING IP', name) - except neutronclient.common.exceptions.NotFound: - self.report_not_found('FLOATING IP', name) - except KeyError: - pass - - try: - for id, name in self.resources['routers'].iteritems(): - try: - if self.dryrun: - self.neutron.show_router(id) - self.report_deletion('Router Gateway', name) - port_list = self.neutron.list_ports(id)['ports'] - for port in port_list: - if 'fixed_ips' in port: - self.report_deletion('Router Interface', - port['fixed_ips'][0]['ip_address']) - else: - self.neutron.remove_gateway_router(id) - self.report_deletion('Router Gateway', name) - # need to delete each interface before deleting the router - port_list = self.neutron.list_ports(id)['ports'] - for port in port_list: - self.remove_router_interface(id, port) - self.neutron.delete_router(id) - self.report_deletion('ROUTER', name) - except neutronclient.common.exceptions.NotFound: - self.report_not_found('ROUTER', name) - except neutronclient.common.exceptions.Conflict as exc: - self.report_error('ROUTER', name, str(exc)) - except KeyError: - pass - try: - for id, name in self.resources['networks'].iteritems(): - try: - if self.dryrun: - self.neutron.show_network(id) - else: - self.remove_network_ports(id) - self.neutron.delete_network(id) - self.report_deletion('NETWORK', name) - except neutronclient.common.exceptions.NetworkNotFoundClient: - self.report_not_found('NETWORK', name) - except neutronclient.common.exceptions.NetworkInUseClient as exc: - self.report_error('NETWORK', name, str(exc)) - except KeyError: - pass - -class KeystoneCleaner(AbstractCleaner): - - def __init__(self, sess, resources, dryrun): - self.keystone = keystoneclient.Client(endpoint_type='publicURL', session=sess) - self.tenant_api = self.keystone.tenants \ - if self.keystone.version == 'v2.0' else self.keystone.projects - res_desc = { - 'users': [self.keystone.users.list], - 'tenants': [self.tenant_api.list] - } - super(KeystoneCleaner, self).__init__('Keystone', res_desc, resources, dryrun) - - def clean(self): - print '*** KEYSTONE cleanup' - try: - for id, name in self.resources['users'].iteritems(): - try: - if self.dryrun: - self.keystone.users.get(id) - else: - self.keystone.users.delete(id) - self.report_deletion('USER', name) - except keystoneclient.auth.exceptions.http.NotFound: - self.report_not_found('USER', name) - except KeyError: - pass - - try: - for id, name in self.resources['tenants'].iteritems(): - try: - if self.dryrun: - self.tenant_api.get(id) - else: - self.tenant_api.delete(id) - self.report_deletion('TENANT', name) - except keystoneclient.auth.exceptions.http.NotFound: - self.report_not_found('TENANT', name) - except KeyError: - pass - -class Cleaners(object): - - def __init__(self, creds_obj, resources, dryrun): - self.cleaners = [] - sess = creds_obj.get_session() - for cleaner_type in [StorageCleaner, ComputeCleaner, NetworkCleaner, KeystoneCleaner]: - self.cleaners.append(cleaner_type(sess, resources, dryrun)) - - def show_resources(self): - table = [["Type", "Name", "UUID"]] - for cleaner in self.cleaners: - table.extend(cleaner.get_resource_list()) - count = len(table) - 1 - print - if count: - print 'SELECTED RESOURCES:' - print tabulate(table, headers="firstrow", tablefmt="psql") - else: - print 'There are no resources to delete.' - print - return count - - def clean(self): - for cleaner in self.cleaners: - cleaner.clean() - -# A dictionary of resources to cleanup -# First level keys are: -# flavors, keypairs, users, routers, floating_ips, instances, volumes, sec_groups, tenants, networks -# second level keys are the resource IDs -# values are the resource name (e.g. 'nfvbench-net0') -def get_resources_from_cleanup_log(logfile): - '''Load triplets separated by '|' into a 2 level dictionary - ''' - resources = {} - with open(logfile) as ff: - content = ff.readlines() - for line in content: - tokens = line.strip().split('|') - restype = tokens[0] - resname = tokens[1] - resid = tokens[2] - if not resid: - # normally only the keypairs have no ID - if restype != "keypairs": - print 'Error: resource type %s has no ID - ignored!!!' % (restype) - else: - resid = '0' - if restype not in resources: - resources[restype] = {} - tres = resources[restype] - tres[resid] = resname - return resources - - -def main(): - parser = argparse.ArgumentParser(description='NFVbench Force Cleanup') - - parser.add_argument('-r', '--rc', dest='rc', - action='store', required=False, - help='openrc file', - metavar='<file>') - parser.add_argument('-f', '--file', dest='file', - action='store', required=False, - help='get resources to delete from cleanup log file ' - '(default:discover from OpenStack)', - metavar='<file>') - parser.add_argument('-d', '--dryrun', dest='dryrun', - action='store_true', - default=False, - help='check resources only - do not delete anything') - parser.add_argument('--filter', dest='filter', - action='store', required=False, - help='resource name regular expression filter (default:"nfvbench")' - ' - OpenStack discovery only', - metavar='<any-python-regex>') - opts = parser.parse_args() - - cred = credentials.Credentials(openrc_file=opts.rc) - - if opts.file: - resources = get_resources_from_cleanup_log(opts.file) - else: - # None means try to find the resources from openstack directly by name - resources = None - global resource_name_re - if opts.filter: - try: - resource_name_re = re.compile(opts.filter) - except Exception as exc: - print 'Provided filter is not a valid python regular expression: ' + opts.filter - print str(exc) - sys.exit(1) - else: - resource_name_re = re.compile('nfvbench') - - cleaners = Cleaners(cred, resources, opts.dryrun) - - if opts.dryrun: - print - print('!!! DRY RUN - RESOURCES WILL BE CHECKED BUT WILL NOT BE DELETED !!!') - print - - # Display resources to be deleted - count = cleaners.show_resources() - if not count: - sys.exit(0) - - if not opts.file and not opts.dryrun: - prompt_to_run() - - cleaners.clean() - -if __name__ == '__main__': - main() diff --git a/docker/Dockerfile b/docker/Dockerfile index bcec2ab..cd59fa0 100644 --- a/docker/Dockerfile +++ b/docker/Dockerfile @@ -2,8 +2,7 @@ FROM ubuntu:16.04 ENV TREX_VER "v2.32" -ENV NFVBENCH_TAG "1.2.2" -ENV VM_IMAGE_VER "0.5" +ENV VM_IMAGE_VER "0.6" RUN apt-get update && apt-get install -y \ git \ @@ -26,9 +25,10 @@ RUN apt-get update && apt-get install -y \ && rm -rf /opt/trex/$TREX_VER/automation/trex_control_plane/stl/trex_stl_lib \ && sed -i -e "s/2048 /512 /" -e "s/2048\"/512\"/" /opt/trex/$TREX_VER/trex-cfg \ && pip install -U pip pbr \ + && hash -r pip \ && pip install -U setuptools \ && cd / \ - && git clone --depth 1 -b $NFVBENCH_TAG https://gerrit.opnfv.org/gerrit/nfvbench \ + && git clone --depth 1 https://gerrit.opnfv.org/gerrit/nfvbench \ && cd /nfvbench && pip install -e . \ && wget -O nfvbenchvm-$VM_IMAGE_VER.qcow2 http://artifacts.opnfv.org/nfvbench/images/nfvbenchvm_centos-$VM_IMAGE_VER.qcow2 \ && python ./docker/cleanup_generators.py \ @@ -39,4 +39,3 @@ RUN apt-get update && apt-get install -y \ ENV TREX_STL_EXT_PATH "/opt/trex/$TREX_VER/external_libs" ENTRYPOINT ["/nfvbench/docker/nfvbench-entrypoint.sh"] - diff --git a/docs/development/design/index.rst b/docs/development/design/index.rst index c54888a..0500ca2 100644 --- a/docs/development/design/index.rst +++ b/docs/development/design/index.rst @@ -12,4 +12,5 @@ OPNFV NFVbench Euphrates Design design versioning + traffic_desc ndrpdr diff --git a/docs/development/design/ndrpdr.rst b/docs/development/design/ndrpdr.rst index 5361174..e34e8ba 100644 --- a/docs/development/design/ndrpdr.rst +++ b/docs/development/design/ndrpdr.rst @@ -6,11 +6,15 @@ NDR/PDR Binary Search ===================== +The NDR/PDR binary search algorithm used by NFVbench is based on the algorithm used by the +FD.io CSIT project, with some additional optimizations. + Algorithm Outline ----------------- -The ServiceChain class is responsible for calculating the NDR/PDR for all frame sizes requested in the configuration. -Calculation for 1 frame size is delegated to the TrafficClient class. +The ServiceChain class (nfvbench/service_chain.py) is responsible for calculating the NDR/PDR +or all frame sizes requested in the configuration. +Calculation for 1 frame size is delegated to the TrafficClient class (nfvbench/traffic_client.py) Call chain for calculating the NDR-PDR for a list of frame sizes: @@ -22,23 +26,58 @@ Call chain for calculating the NDR-PDR for a list of frame sizes: - TrafficClient.__range_search() recursive binary search The search range is delimited by a left and right rate (expressed as a % of line rate per direction). +The search always start at line rate per port, e.g. in the case of 2x10Gbps, the first iteration +will send 10Gbps of traffic on each port. The load_epsilon configuration parameter defines the accuracy of the result as a % of line rate. The default value of 0.1 indicates for example that the measured NDR and PDR are within 0.1% of line rate of the actual NDR/PDR (e.g. 0.1% of 10Gbps is 10Mbps). It also determines how small the search range must be in the binary search. +Smaller values of load_epsilon will result in more iterations and will take more time but may not +always be beneficial if the absolute value falls below the precision level of the measurement. +For example a value of 0.01% would translate to an absolute value of 1Mbps (for a 10Gbps port) or +around 10kpps (at 64 byte size) which might be too fine grain. The recursion narrows down the range by half and stops when: - the range is smaller than the configured load_epsilon value - or when the search hits 100% or 0% of line rate +Optimization +------------ + +Binary search algorithms assume that the drop rate curve is monotonically increasing with the Tx rate. +To save time, the algorithm used by NFVbench is capable of calculating the optimal Tx rate for an +arbitrary list of target maximum drop rates in one pass instead of the usual 1 pass per target maximum drop rate. +This saves time linearly to the number target drop rates. +For example, a typical NDR/PDR search will have 2 target maximum drop rates: + +- NDR = 0.001% +- PDR = 0.1% + +The binary search will then start with a sorted list of 2 target drop rates: [0.1, 0.001]. +The first part of the binary search will then focus on finding the optimal rate for the first target +drop rate (0.1%). When found, the current target drop rate is removed from the list and +iteration continues with the next target drop rate in the list but this time +starting from the upper/lower range of the previous target drop rate, which saves significant time. +The binary search continues until the target maximum drop rate list is empty. + +Results Granularity +------------------- +The binary search results contain per direction stats (forward and reverse). +In the case of multi-chaining, results contain per chain stats. +The current code only reports aggregated stats (forward + reverse for all chains) but could be enhanced +to report per chain stats. + + +CPU Limitations +--------------- One particularity of using a software traffic generator is that the requested Tx rate may not always be met due to resource limitations (e.g. CPU is not fast enough to generate a very high load). The algorithm should take this into consideration: -- always monitor the actual Tx rate achieved +- always monitor the actual Tx rate achieved as reported back by the traffic generator - actual Tx rate is always <= requested Tx rate - the measured drop rate should always be relative to the actual Tx rate -- if the actual Tx rate is < requested Tx rate and the measured drop rate is already within threshold (<NDR/PDR threshold) then the binary search must stop with proper warning - - +- if the actual Tx rate is < requested Tx rate and the measured drop rate is already within threshold + (<NDR/PDR threshold) then the binary search must stop with proper warning because the actual NDR/PDR + might probably be higher than the reported values diff --git a/docs/development/design/traffic_desc.rst b/docs/development/design/traffic_desc.rst new file mode 100644 index 0000000..2a40b6a --- /dev/null +++ b/docs/development/design/traffic_desc.rst @@ -0,0 +1,85 @@ +.. This work is licensed under a Creative Commons Attribution 4.0 International +.. License. +.. http://creativecommons.org/licenses/by/4.0 +.. (c) Cisco Systems, Inc + +Traffic Description +=================== + +The general packet path model followed by NFVbench requires injecting traffic into an arbitrary +number of service chains, where each service chain is identified by 2 edge networks (left and right). +In the current multi-chaining model: + +- all service chains share the same left and right edge networks +- each port associated to the traffic generator is dedicated to send traffic to one edge network + +In an OpenStack deployment, this corresponds to all chains sharing the same 2 neutron networks. +If VLAN encapsulation is used, all traffic sent to a port will have the same VLAN id. + +Basic Packet Description +------------------------ + +The code to create the UDP packet is located in TRex.create_pkt() (nfvbench/traffic_gen/trex.py). + +NFVbench always generates UDP packets (even when doing L2 forwarding). +The final size of the frame containing each UDP packet will be based on the requested L2 frame size. +When taking into account the minimum payload size requirements from the traffic generator for +the latency streams, the minimum L2 frame size is 64 byte (no vlan tagging) or +68 bytes (with vlan tagging). + +Flows Specification +------------------- + +Mac Addresses +............. +The source MAC address is always the local port MAC address (for each port). +The destination MAC address is based on the configuration and can be: + +- the traffic generator peer port MAC address in the case of L2 loopback at the switch level + or when using a loopback cable +- the dest MAC as specified by the configuration file (EXT chain no ARP) +- the dest MAC as discovered by ARP (EXT chain) +- the VM MAC as dicovered from Neutron API (PVP, PVVP chains) + +NFVbench does not currently range on the MAC addresses. + +IP addresses +............ +The source IP address is fixed per chain. +The destination IP address is variable within a distinct range per chain. + +UDP ports +......... +The source and destination ports are fixed for all packets and can be set in the configuratoon +file (default is 53). + +Payload User Data +................. +The length of the user data is based on the requested L2 frame size and takes into account the +size of the L2 header - including the VLAN tag if applicable. + + +IMIX Support +------------ +In the case of IMIX, each direction is made of 4 streams: +- 1 latency stream +- 1 stream for each IMIX frame size + +The IMIX ratio is encoded into the number of consecutive packets sent by each stream in turn. + +Service Chains and Streams +-------------------------- +A stream identifies one "stream" of packets with same characteristics such as rate and destination address. +NFVbench will create 2 streams per service chain per direction: + +- 1 latency stream set to 1000pps +- 1 main traffic stream set to the requested Tx rate less the latency stream rate (1000pps) + +For example, a benchmark with 1 chain (fixed rate) will result in a total of 4 streams. +A benchmark with 20 chains will results in a total of 80 streams (fixed rate, it is more with IMIX). + +The overall flows are split equally between the number of chains by using the appropriate destination +MAC address. + +For example, in the case of 10 chains, 1M flows and fixed rate, there will be a total of 40 streams. +Each of the 20 non-latency stream will generate packets corresponding to 50,000 flows (unique src/dest address tuples). diff --git a/docs/release/release-notes/release-notes.rst b/docs/release/release-notes/release-notes.rst index a3402cb..7c9cbcb 100644 --- a/docs/release/release-notes/release-notes.rst +++ b/docs/release/release-notes/release-notes.rst @@ -2,6 +2,30 @@ .. http://creativecommons.org/licenses/by/4.0 .. (c) Cisco Systems, Inc +RELEASE NOTES ++++++++++++++ + +OPNFV Fraser Release +==================== + +Over 30 Jira tickets have been addressed in this release (Jira NFVBENCH-55 to NFVBENCH-78) + +The Fraser release adds the following new features: + +- support for benchmarking non-OpenStack environments (with external setup and no OpenStack openrc file) +- PVVP packet path with SRIOV at the edge and vswitch between VMs +- support logging events and results through fluentd + +Enhancements and main bug fixes: + +- end to end connectivity for larger chain count is now much more accurate for large chain count - avoiding excessive drops +- use newer version of TRex (2.32) +- use newer version of testpmd DPDK +- NDR/PDR uses actual TX rate to calculate drops - resulting in more accurate results +- add pylint to unit testing +- add self sufficient and standalone unit testing (without actual testbed) + + OPNFV Euphrates Release ======================= diff --git a/docs/testing/user/userguide/advanced.rst b/docs/testing/user/userguide/advanced.rst index 252cbc9..02c7fce 100644 --- a/docs/testing/user/userguide/advanced.rst +++ b/docs/testing/user/userguide/advanced.rst @@ -314,46 +314,60 @@ NFVbench will dicover the MAC addresses to use for generated frames using: - either OpenStack discovery (find the MAC of an existing VM) in the case of PVP and PVVP service chains - or using dynamic ARP discovery (find MAC from IP) in the case of external chains. -Cleanup Script --------------- +Status and Cleanup of NFVbench Resources +---------------------------------------- + +The --status option will display the status of NFVbench and list any NFVbench resources. You need to pass the OpenStack RC +file in order to connect to OpenStack. + +.. code-block:: none -The nfvbench_cleanup script will cleanup resources created by NFVbench. You need to pass the OpenStack RC file in order to connect to -OpenStack. + # nfvbench --status -r /tmp/nfvbench/openrc + 2018-04-09 17:05:48,682 INFO Version: 1.3.2.dev1 + 2018-04-09 17:05:48,683 INFO Status: idle + 2018-04-09 17:05:48,757 INFO Discovering instances nfvbench-loop-vm... + 2018-04-09 17:05:49,252 INFO Discovering flavor nfvbench.medium... + 2018-04-09 17:05:49,281 INFO Discovering networks... + 2018-04-09 17:05:49,365 INFO No matching NFVbench resources found + # + +The Status can be either "idle" or "busy (run pending)". + +The --cleanup option will first discover resources created by NFVbench and prompt if you want to proceed with cleaning them up. Example of run: .. code-block:: none - # nfvbench_cleanup -r /tmp/nfvbench/openrc - Discovering Storage resources... - Discovering Compute resources... - Discovering Network resources... - Discovering Keystone resources... - - SELECTED RESOURCES: - +-----------+-------------------+--------------------------------------+ - | Type | Name | UUID | - |-----------+-------------------+--------------------------------------| - | flavors | nfvbench.medium | 362b2215-89d1-4f46-8b89-8e58165ff5bc | - | instances | nfvbench-loop-vm0 | f78dfb74-1b8e-4c5c-8d83-652a7571da95 | - | networks | nfvbench-net0 | 57d7e6c9-325f-4c13-9b1b-929344cc9c39 | - | networks | nfvbench-net1 | 2d429bcd-33fa-4aa4-9f2e-299a735177c9 | - +-----------+-------------------+--------------------------------------+ - - Warning: You didn't specify a resource list file as the input. The script will delete all resources shown above. + # nfvbench --cleanup -r /tmp/nfvbench/openrc + 2018-04-09 16:58:00,204 INFO Version: 1.3.2.dev1 + 2018-04-09 16:58:00,205 INFO Status: idle + 2018-04-09 16:58:00,279 INFO Discovering instances nfvbench-loop-vm... + 2018-04-09 16:58:00,829 INFO Discovering flavor nfvbench.medium... + 2018-04-09 16:58:00,876 INFO Discovering networks... + 2018-04-09 16:58:00,960 INFO Discovering ports... + 2018-04-09 16:58:01,012 INFO Discovered 6 NFVbench resources: + +----------+-------------------+--------------------------------------+ + | Type | Name | UUID | + |----------+-------------------+--------------------------------------| + | Instance | nfvbench-loop-vm0 | b039b858-777e-467e-99fb-362f856f4a94 | + | Flavor | nfvbench.medium | a027003c-ad86-4f24-b676-2b05bb06adc0 | + | Network | nfvbench-net0 | bca8d183-538e-4965-880e-fd92d48bfe0d | + | Network | nfvbench-net1 | c582a201-8279-4309-8084-7edd6511092c | + | Port | | 67740862-80ac-4371-b04e-58a0b0f05085 | + | Port | | b5db95b9-e419-4725-951a-9a8f7841e66a | + +----------+-------------------+--------------------------------------+ + 2018-04-09 16:58:01,013 INFO NFVbench will delete all resources shown... Are you sure? (y/n) y - *** STORAGE cleanup - *** COMPUTE cleanup - . Waiting for 1 instances to be fully deleted... - . INSTANCE 1 left to be deleted, retries left=5... - . INSTANCE 1 left to be deleted, retries left=4... - + INSTANCE nfvbench-loop-vm0 is successfully deleted - + FLAVOR nfvbench.medium is successfully deleted - *** NETWORK cleanup - + Network port 075d91f3-fa6a-428c-bd3f-ebd40cd935e1 is successfully deleted - + Network port 3a7ccd8c-53a6-43d0-a823-4b5ca762d06e is successfully deleted - + NETWORK nfvbench-net0 is successfully deleted - + Network port 5b5a75bd-e0b5-4f81-91b9-9e216d194f48 is successfully deleted - + Network port cc2d8f1b-49fe-491e-9e44-6990fc57e891 is successfully deleted - + NETWORK nfvbench-net1 is successfully deleted - *** KEYSTONE cleanup + 2018-04-09 16:58:01,865 INFO Deleting instance nfvbench-loop-vm0... + 2018-04-09 16:58:02,058 INFO Waiting for 1 instances to be fully deleted... + 2018-04-09 16:58:02,182 INFO 1 yet to be deleted by Nova, retries left=6... + 2018-04-09 16:58:04,506 INFO 1 yet to be deleted by Nova, retries left=5... + 2018-04-09 16:58:06,636 INFO 1 yet to be deleted by Nova, retries left=4... + 2018-04-09 16:58:08,701 INFO Deleting flavor nfvbench.medium... + 2018-04-09 16:58:08,729 INFO Deleting port 67740862-80ac-4371-b04e-58a0b0f05085... + 2018-04-09 16:58:09,102 INFO Deleting port b5db95b9-e419-4725-951a-9a8f7841e66a... + 2018-04-09 16:58:09,620 INFO Deleting network nfvbench-net0... + 2018-04-09 16:58:10,357 INFO Deleting network nfvbench-net1... # + +The --force-cleanup option will do the same but without prompting for confirmation. diff --git a/docs/testing/user/userguide/quickstart_docker.rst b/docs/testing/user/userguide/quickstart_docker.rst index c5e5eda..a5152cf 100644 --- a/docs/testing/user/userguide/quickstart_docker.rst +++ b/docs/testing/user/userguide/quickstart_docker.rst @@ -115,7 +115,7 @@ To run NFVBench enabling REST server (mount the configuration json and the path .. code-block:: bash cd ~/nfvbench_ws - docker run --detach --net=host --privileged -e HOST="127.0.0.1" -e PORT=7556 --e CONFIG_FILE="/tmp/nfvbench/nfvbenchconfig.json -v $PWD:/tmp/nfvbench -v /dev:/dev -v /lib/modules/$(uname -r):/lib/modules/$(uname -r) -v /usr/src/kernels:/usr/src/kernels --name nfvbench opnfv/nfvbench start_rest_server + docker run --detach --net=host --privileged -e HOST="127.0.0.1" -e PORT=7556 -e CONFIG_FILE="/tmp/nfvbench/nfvbenchconfig.json -v $PWD:/tmp/nfvbench -v /dev:/dev -v /lib/modules/$(uname -r):/lib/modules/$(uname -r) -v /usr/src/kernels:/usr/src/kernels --name nfvbench opnfv/nfvbench start_rest_server The create an alias to make it easy to execute nfvbench commands directly from the host shell prompt: diff --git a/docs/testing/user/userguide/readme.rst b/docs/testing/user/userguide/readme.rst index 785ffed..b437ff9 100644 --- a/docs/testing/user/userguide/readme.rst +++ b/docs/testing/user/userguide/readme.rst @@ -46,11 +46,21 @@ main purpose is to measure the performance of the NFVi infrastructure which is m External Chains --------------- -NFVbench also supports settings that involve externally staged packet paths with or without OpenStack: +NFVbench supports settings that involve externally staged packet paths with or without OpenStack: - run benchmarks on existing service chains at the L3 level that are staged externally by any other tool (e.g. any VNF capable of L3 routing) - run benchmarks on existing L2 chains that are configured externally (e.g. pure L2 forwarder such as DPDK testpmd) +Direct L2 Loopback (Switch or wire loopback) +-------------------------------------------- +NFVbench supports benchmarking of pure L2 loopbacks (see "--l2-loopback vlan" option) + +- Switch level loopback +- Port to port wire loopback + +In this mode, NFVbench will take a vlan ID and send packets from each port to the other port +(dest MAC set to the other port MAC) using the same VLAN ID on both ports. +This can be useful for example to verify that the connectivity to the switch is working properly. Traffic Generation ------------------ @@ -116,7 +126,8 @@ Multi-Chaining (N*PVP or N*PVVP) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Multiple service chains can be setup by NFVbench without any limit on the concurrency (other than limits imposed by available resources on compute nodes). -In the case of multiple service chains, NFVbench will instruct the traffic generator to use multiple L3 packet streams (frames directed to each path will have a unique destination MAC address). +In the case of multiple service chains, NFVbench will instruct the traffic generator to use multiple L3 packet streams (frames directed to each path will +have a unique destination MAC address). Example of multi-chaining with 2 concurrent PVP service chains: @@ -126,6 +137,8 @@ This innovative feature will allow to measure easily the performance of a fully Multi-chaining is currently limited to 1 compute node (PVP or PVVP intra-node) or 2 compute nodes (for PVVP inter-node). The 2 edge interfaces for all service chains will share the same 2 networks. +The total traffic will be split equally across all chains. + SR-IOV ^^^^^^ @@ -179,7 +192,7 @@ NFVbench is agnostic of the virtual switch implementation and has been tested wi - - - - +Limitations +*********** +NFVbench only supports VLAN with OpenStack. +NFVbench does not support VxLAN overlays. diff --git a/nfvbench/cfg.default.yaml b/nfvbench/cfg.default.yaml index e1c05c3..2af6d63 100755 --- a/nfvbench/cfg.default.yaml +++ b/nfvbench/cfg.default.yaml @@ -131,6 +131,11 @@ sriov: false # Can be overriden by --no-int-config no_int_config: false +# Perform port to port loopback (direct or through switch) +# Should be used with EXT service chain and no ARP (no_arp: true) +# Can be overriden by --l2-loopback +l2_loopback: false + # Resources created by NFVbench will not be removed # Can be overriden by --no-cleanup no_cleanup: false @@ -231,8 +236,8 @@ loop_vm_name: 'nfvbench-loop-vm' # - PVVP uses left, middle and right # - for EXT chains, this structure is not relevant - refer to external_networks # Otherwise a new internal network will be created with that name, subnet and CIDR. -# -# segmentation_id can be set to enforce a specific VLAN id - by default (empty) the VLAN id +# +# segmentation_id can be set to enforce a specific VLAN id - by default (empty) the VLAN id # will be assigned by Neutron. # Must be unique for each network # physical_network can be set to pick a specific phsyical network - by default (empty) the @@ -284,14 +289,14 @@ external_networks: right: 'nfvbench-net1' # Use 'true' to enable VLAN tagging of packets generated and sent by the traffic generator -# Leave empty you do not want the traffic generator to insert the VLAN tag. This is +# Leave empty you do not want the traffic generator to insert the VLAN tag. This is # needed for example if VLAN tagging is enabled on switch (trunk mode) or if you want to hook directly to a NIC # By default is set to true (which is the nominal use case with TOR and trunk mode to Trex) vlan_tagging: true # Specify only when you want to override VLAN IDs used for tagging with own values (exactly 2). -# Default behavior of VLAN tagging is to retrieve VLAN IDs from OpenStack networks provided above. -# In case of VxLAN this setting is ignored and only vtep_vlan from traffic generator profile is used. +# Default behavior (empty list) is to retrieve VLAN IDs from OpenStack networks described in external_networks. +# This property is ignored in the case of l2-loopback # Example: [1998, 1999] vlans: [] @@ -354,6 +359,9 @@ duration_sec: 60 # Can be overridden by --interval interval_sec: 10 +# Default pause between iterations of a binary search (NDR/PDR) +pause_sec: 2 + # NDR / PDR configuration measurement: # Drop rates represent the ratio of dropped packet to the total number of packets sent. diff --git a/nfvbench/chain_clients.py b/nfvbench/chain_clients.py index faf7c2a..71c6c97 100644 --- a/nfvbench/chain_clients.py +++ b/nfvbench/chain_clients.py @@ -185,7 +185,7 @@ class BasicStageClient(object): return None return availability_zone + ':' + host - def _lookup_servers(self, name=None, nets=None, az=None, flavor_id=None): + def _lookup_servers(self, name=None, nets=None, flavor_id=None): error_msg = 'VM with the same name, but non-matching {} found. Aborting.' networks = set([net['name'] for net in nets]) if nets else None server_list = self.comp.get_server_list() @@ -224,7 +224,10 @@ class BasicStageClient(object): files={nfvbenchvm_config_location: nfvbenchvm_config}) if server: setattr(server, 'is_reuse', False) - LOG.info('Creating instance: %s on %s', name, az) + msg = 'Creating instance: %s' % name + if az: + msg += ' on %s' % az + LOG.info(msg) else: raise StageClientException('Unable to create instance: %s.' % (name)) return server @@ -380,8 +383,8 @@ class BasicStageClient(object): compute_nodes.append(az + ':' + hostname) return compute_nodes - def get_reusable_vm(self, name, nets, az): - servers = self._lookup_servers(name=name, nets=nets, az=az, + def get_reusable_vm(self, name, nets): + servers = self._lookup_servers(name=name, nets=nets, flavor_id=self.flavor_type['flavor'].id) if servers: server = servers[0] @@ -477,17 +480,19 @@ class PVPStageClient(BasicStageClient): nets = self.config.internal_networks self.nets.extend([self._create_net(**n) for n in [nets.left, nets.right]]) - az_list = self.comp.get_enabled_az_host_list(required_count=1) - if not az_list: - raise Exception('Not enough hosts found.') + if self.comp.config.compute_nodes: + az_list = self.comp.get_enabled_az_host_list(required_count=1) + if not az_list: + raise Exception('Not enough hosts found.') + az = az_list[0] + else: + az = None - az = az_list[0] - self.compute_nodes.add(az) for chain_index in xrange(self.config.service_chain_count): name = self.config.loop_vm_name + str(chain_index) - reusable_vm = self.get_reusable_vm(name, self.nets, az) - if reusable_vm: - self.vms.append(reusable_vm) + server = self.get_reusable_vm(name, self.nets) + if server: + self.vms.append(server) else: vnic_type = 'direct' if self.config.sriov else 'normal' ports = [self._create_port(net, vnic_type) for net in self.nets] @@ -497,8 +502,19 @@ class PVPStageClient(BasicStageClient): ports[0]['mac_address'], ports[1]['mac_address']) self.created_ports.extend(ports) - self.vms.append(self._create_server(name, ports, az, config_file)) + server = self._create_server(name, ports, az, config_file) + self.vms.append(server) + + if chain_index == 0: + # First VM, save the hypervisor name. Used in future for + # maintain affinity. + self._ensure_vms_active() + server = self.comp.poll_server(server) + az = "%s:%s" % (getattr(server, 'OS-EXT-AZ:availability_zone'), + getattr(server, 'OS-EXT-SRV-ATTR:hypervisor_hostname')) + self._ensure_vms_active() + self.compute_nodes = set(self.get_loop_vm_compute_nodes()) self.set_ports() @@ -519,37 +535,36 @@ class PVVPStageClient(BasicStageClient): nets = self.config.internal_networks self.nets.extend([self._create_net(**n) for n in [nets.left, nets.right, nets.middle]]) - required_count = 2 if self.config.inter_node else 1 - az_list = self.comp.get_enabled_az_host_list(required_count=required_count) - - if not az_list: - raise Exception('Not enough hosts found.') - - az1 = az2 = az_list[0] - if self.config.inter_node: - if len(az_list) > 1: - az1 = az_list[0] - az2 = az_list[1] - else: - # fallback to intra-node - az1 = az2 = az_list[0] - self.config.inter_node = False - LOG.info('Using intra-node instead of inter-node.') - - self.compute_nodes.add(az1) - self.compute_nodes.add(az2) + if self.comp.config.compute_nodes: + required_count = 2 if self.config.inter_node else 1 + az_list = self.comp.get_enabled_az_host_list(required_count=required_count) + if not az_list: + raise Exception('Not enough hosts found.') + + az1 = az2 = az_list[0] + if self.config.inter_node: + if len(az_list) > 1: + az1 = az_list[0] + az2 = az_list[1] + else: + # fallback to intra-node + az1 = az2 = az_list[0] + self.config.inter_node = False + LOG.info('Using intra-node instead of inter-node.') + else: + az1 = az2 = None # Create loop VMs for chain_index in xrange(self.config.service_chain_count): name0 = self.config.loop_vm_name + str(chain_index) + 'a' # Attach first VM to net0 and net2 vm0_nets = self.nets[0::2] - reusable_vm0 = self.get_reusable_vm(name0, vm0_nets, az1) + reusable_vm0 = self.get_reusable_vm(name0, vm0_nets) name1 = self.config.loop_vm_name + str(chain_index) + 'b' # Attach second VM to net1 and net2 vm1_nets = self.nets[1:] - reusable_vm1 = self.get_reusable_vm(name1, vm1_nets, az2) + reusable_vm1 = self.get_reusable_vm(name1, vm1_nets) if reusable_vm0 and reusable_vm1: self.vms.extend([reusable_vm0, reusable_vm1]) @@ -583,14 +598,36 @@ class PVVPStageClient(BasicStageClient): vm1_port_net2['mac_address'], vm1_port_net1['mac_address']) - self.vms.append(self._create_server(name0, - [vm0_port_net0, vm0_port_net2], - az1, - config_file0)) - self.vms.append(self._create_server(name1, - [vm1_port_net2, vm1_port_net1], - az2, - config_file1)) + vm1 = self._create_server(name0, [vm0_port_net0, vm0_port_net2], az1, config_file0) + self.vms.append(vm1) + if chain_index == 0: + # First VM on first chain, save the hypervisor name. Used + # in future for maintain affinity. + self._ensure_vms_active() + vm1 = self.comp.poll_server(vm1) + az1 = "%s:%s" % (getattr(vm1, 'OS-EXT-AZ:availability_zone'), + getattr(vm1, 'OS-EXT-SRV-ATTR:hypervisor_hostname')) + if not self.config.inter_node: + # By default, NOVA scheduler will try first with + # different hypervisor for workload balance, but when + # inter-node is not configured, use the same AZ to run + # intra-node test case. + az2 = az1 + + vm2 = self._create_server(name1, [vm1_port_net2, vm1_port_net1], az2, config_file1) + self.vms.append(vm2) + if chain_index == 0 and self.config.inter_node: + # Second VM on first chain, save the hypervisor name. Used + # in future for maintain affinity. + self._ensure_vms_active() + vm2 = self.comp.poll_server(vm2) + az2 = "%s:%s" % (getattr(vm2, 'OS-EXT-AZ:availability_zone'), + getattr(vm2, 'OS-EXT-SRV-ATTR:hypervisor_hostname')) + if az1 == az2: + # Configure to run inter-node, but not enough node to run + self.config.inter_node = False + LOG.info('Using intra-node instead of inter-node.') self._ensure_vms_active() + self.compute_nodes = set(self.get_loop_vm_compute_nodes()) self.set_ports() diff --git a/nfvbench/chain_managers.py b/nfvbench/chain_managers.py index 087c751..9cd6c7d 100644 --- a/nfvbench/chain_managers.py +++ b/nfvbench/chain_managers.py @@ -15,7 +15,6 @@ # import time - from log import LOG from network import Network from packet_analyzer import PacketAnalyzer @@ -24,6 +23,7 @@ from stats_collector import IntervalCollector class StageManager(object): + """A class to stage resources in the systenm under test.""" def __init__(self, config, cred, factory): self.config = config @@ -56,7 +56,7 @@ class StageManager(object): return self.client.ports def get_compute_nodes(self): - return self.client.compute_nodes + return self.client.compute_nodes if self.client else {} def set_vm_macs(self): if self.client and self.config.service_chain != ChainType.EXT: @@ -67,7 +67,8 @@ class StageManager(object): self.client.dispose() -class StatsManager(object): +class PVPStatsManager(object): + """A class to generate traffic and extract results for PVP chains.""" def __init__(self, config, clients, specs, factory, vlans, notifier=None): self.config = config @@ -89,19 +90,31 @@ class StatsManager(object): try: self.worker.set_vlans(self.vlans) self._config_interfaces() - except Exception as exc: + except Exception: # since the wrorker is up and running, we need to close it # in case of exception self.close() - raise exc + raise def _get_data(self): return self.worker.get_data() if self.worker else {} - def _get_network(self, traffic_port, index=None, reverse=False): - interfaces = [self.clients['traffic'].get_interface(traffic_port)] + def _get_network(self, traffic_port, stats, reverse=False): + """Get the Network object corresponding to a given TG port. + + :param traffic_port: must be either 0 or 1 + :param stats: TG stats for given traffic port + :param reverse: specifies if the interface list for this network + should go from TG to loopback point (reverse=false) or + from loopback point to TG (reverse=true) + """ + # build the interface list in fwd direction (TG To loopback point) + interfaces = [self.clients['traffic'].get_interface(traffic_port, stats)] if self.worker: - interfaces.extend(self.worker.get_network_interfaces(index)) + # if available, + # interfaces for workers must be aligned on the TG port number + interfaces.extend(self.worker.get_network_interfaces(traffic_port)) + # let Network reverse the interface order if needed return Network(interfaces, reverse) def _config_interfaces(self): @@ -131,9 +144,7 @@ class StatsManager(object): return self.worker.get_version() if self.worker else {} def run(self): - """ - Run analysis in both direction and return the analysis - """ + """Run analysis in both direction and return the analysis.""" if self.worker: self.worker.run() @@ -144,16 +155,18 @@ class StatsManager(object): 'stats': stats } + # fetch latest stats from traffic gen + stats = self.clients['traffic'].get_stats() LOG.info('Requesting packet analysis on the forward direction...') result['packet_analysis']['direction-forward'] = \ - self.get_analysis([self._get_network(0, 0), - self._get_network(0, 1, reverse=True)]) + self.get_analysis([self._get_network(0, stats), + self._get_network(1, stats, reverse=True)]) LOG.info('Packet analysis on the forward direction completed') LOG.info('Requesting packet analysis on the reverse direction...') result['packet_analysis']['direction-reverse'] = \ - self.get_analysis([self._get_network(1, 1), - self._get_network(1, 0, reverse=True)]) + self.get_analysis([self._get_network(1, stats), + self._get_network(0, stats, reverse=True)]) LOG.info('Packet analysis on the reverse direction completed') return result @@ -182,21 +195,14 @@ class StatsManager(object): self.worker.close() -class PVPStatsManager(StatsManager): +class PVVPStatsManager(PVPStatsManager): + """A Class to generate traffic and extract results for PVVP chains.""" def __init__(self, config, clients, specs, factory, vlans, notifier=None): - StatsManager.__init__(self, config, clients, specs, factory, vlans, notifier) - - -class PVVPStatsManager(StatsManager): - - def __init__(self, config, clients, specs, factory, vlans, notifier=None): - StatsManager.__init__(self, config, clients, specs, factory, vlans, notifier) + PVPStatsManager.__init__(self, config, clients, specs, factory, vlans, notifier) def run(self): - """ - Run analysis in both direction and return the analysis - """ + """Run analysis in both direction and return the analysis.""" fwd_v2v_net, rev_v2v_net = self.worker.run() stats = self._generate_traffic() @@ -205,16 +211,17 @@ class PVVPStatsManager(StatsManager): 'packet_analysis': {}, 'stats': stats } - - fwd_nets = [self._get_network(0, 0)] + # fetch latest stats from traffic gen + stats = self.clients['traffic'].get_stats() + fwd_nets = [self._get_network(0, stats)] if fwd_v2v_net: fwd_nets.append(fwd_v2v_net) - fwd_nets.append(self._get_network(0, 1, reverse=True)) + fwd_nets.append(self._get_network(1, stats, reverse=True)) - rev_nets = [self._get_network(1, 1)] + rev_nets = [self._get_network(1, stats)] if rev_v2v_net: rev_nets.append(rev_v2v_net) - rev_nets.append(self._get_network(1, 0, reverse=True)) + rev_nets.append(self._get_network(0, stats, reverse=True)) LOG.info('Requesting packet analysis on the forward direction...') result['packet_analysis']['direction-forward'] = self.get_analysis(fwd_nets) @@ -227,9 +234,11 @@ class PVVPStatsManager(StatsManager): return result -class EXTStatsManager(StatsManager): +class EXTStatsManager(PVPStatsManager): + """A Class to generate traffic and extract results for EXT chains.""" + def __init__(self, config, clients, specs, factory, vlans, notifier=None): - StatsManager.__init__(self, config, clients, specs, factory, vlans, notifier) + PVPStatsManager.__init__(self, config, clients, specs, factory, vlans, notifier) def _setup(self): if self.specs.openstack: diff --git a/nfvbench/cleanup.py b/nfvbench/cleanup.py new file mode 100644 index 0000000..246be3f --- /dev/null +++ b/nfvbench/cleanup.py @@ -0,0 +1,179 @@ +#!/usr/bin/env python +# Copyright 2017 Cisco Systems, Inc. All rights reserved. +# +# Licensed under the Apache License, Version 2.0 (the "License"); you may +# not use this file except in compliance with the License. You may obtain +# a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT +# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the +# License for the specific language governing permissions and limitations +# under the License. +# + +import sys +import time + +from neutronclient.neutron import client as nclient +from novaclient.client import Client +from novaclient.exceptions import NotFound +from tabulate import tabulate + +import credentials as credentials +from log import LOG + +class ComputeCleaner(object): + """A cleaner for compute resources.""" + + def __init__(self, nova_client, instance_prefix): + self.nova_client = nova_client + LOG.info('Discovering instances %s...', instance_prefix) + all_servers = self.nova_client.servers.list() + self.servers = [server for server in all_servers + if server.name.startswith(instance_prefix)] + + def instance_exists(self, server): + try: + self.nova_client.servers.get(server.id) + except NotFound: + return False + return True + + def get_resource_list(self): + return [["Instance", server.name, server.id] for server in self.servers] + + def clean(self): + if self.servers: + for server in self.servers: + try: + LOG.info('Deleting instance %s...', server.name) + self.nova_client.servers.delete(server.id) + except Exception: + LOG.exception("Instance %s deletion failed", server.name) + LOG.info(' Waiting for %d instances to be fully deleted...', len(self.servers)) + retry_count = 5 + len(self.servers) * 2 + while True: + retry_count -= 1 + self.servers = [server for server in self.servers if self.instance_exists(server)] + if not self.servers: + break + + if retry_count: + LOG.info(' %d yet to be deleted by Nova, retries left=%d...', + len(self.servers), retry_count) + time.sleep(2) + else: + LOG.warning(' instance deletion verification timed out: %d not removed', + len(self.servers)) + break + + +class NetworkCleaner(object): + """A cleaner for network resources.""" + + def __init__(self, neutron_client, network_names): + self.neutron_client = neutron_client + LOG.info('Discovering networks...') + all_networks = self.neutron_client.list_networks()['networks'] + self.networks = [] + for net in all_networks: + try: + network_names.remove(net['name']) + self.networks.append(net) + except ValueError: + pass + if not network_names: + break + net_ids = [net['id'] for net in self.networks] + if net_ids: + LOG.info('Discovering ports...') + all_ports = self.neutron_client.list_ports()['ports'] + self.ports = [port for port in all_ports if port['network_id'] in net_ids] + else: + self.ports = [] + + def get_resource_list(self): + res_list = [["Network", net['name'], net['id']] for net in self.networks] + res_list.extend([["Port", port['name'], port['id']] for port in self.ports]) + return res_list + + def clean(self): + for port in self.ports: + LOG.info("Deleting port %s...", port['id']) + try: + self.neutron_client.delete_port(port['id']) + except Exception: + LOG.exception("Port deletion failed") + + for net in self.networks: + LOG.info("Deleting network %s...", net['name']) + try: + self.neutron_client.delete_network(net['id']) + except Exception: + LOG.exception("Network deletion failed") + +class FlavorCleaner(object): + """Cleaner for NFVbench flavor.""" + + def __init__(self, nova_client, name): + self.name = name + LOG.info('Discovering flavor %s...', name) + try: + self.flavor = nova_client.flavors.find(name=name) + except NotFound: + self.flavor = None + + def get_resource_list(self): + if self.flavor: + return [['Flavor', self.name, self.flavor.id]] + return None + + def clean(self): + if self.flavor: + LOG.info("Deleting flavor %s...", self.flavor.name) + try: + self.flavor.delete() + except Exception: + LOG.exception("Flavor deletion failed") + +class Cleaner(object): + """Cleaner for all NFVbench resources.""" + + def __init__(self, config): + cred = credentials.Credentials(config.openrc_file, None, False) + session = cred.get_session() + self.neutron_client = nclient.Client('2.0', session=session) + self.nova_client = Client(2, session=session) + network_names = [inet['name'] for inet in config.internal_networks.values()] + self.cleaners = [ComputeCleaner(self.nova_client, config.loop_vm_name), + FlavorCleaner(self.nova_client, config.flavor_type), + NetworkCleaner(self.neutron_client, network_names)] + + def show_resources(self): + """Show all NFVbench resources.""" + table = [["Type", "Name", "UUID"]] + for cleaner in self.cleaners: + res_list = cleaner.get_resource_list() + if res_list: + table.extend(res_list) + count = len(table) - 1 + if count: + LOG.info('Discovered %d NFVbench resources:', count) + print tabulate(table, headers="firstrow", tablefmt="psql") + else: + LOG.info('No matching NFVbench resources found') + return count + + def clean(self, prompt): + """Clean all resources.""" + LOG.info("NFVbench will delete all resources shown...") + if prompt: + answer = raw_input("Are you sure? (y/n) ") + if answer.lower() != 'y': + LOG.info("Exiting without deleting any resource") + sys.exit(0) + for cleaner in self.cleaners: + cleaner.clean() diff --git a/nfvbench/config_plugin.py b/nfvbench/config_plugin.py index 78c2ebb..f6654eb 100644 --- a/nfvbench/config_plugin.py +++ b/nfvbench/config_plugin.py @@ -49,7 +49,7 @@ class ConfigPluginBase(object): """Returns OpenStack specs for host.""" @abc.abstractmethod - def get_run_spec(self, openstack_spec): + def get_run_spec(self, config, openstack_spec): """Returns RunSpec for given platform.""" @abc.abstractmethod @@ -81,9 +81,9 @@ class ConfigPlugin(ConfigPluginBase): """Returns OpenStack specs for host.""" return specs.OpenStackSpec() - def get_run_spec(self, openstack_spec): + def get_run_spec(self, config, openstack_spec): """Returns RunSpec for given platform.""" - return specs.RunSpec(self.config.no_vswitch_access, openstack_spec) + return specs.RunSpec(config.no_vswitch_access, openstack_spec) def validate_config(self, config, openstack_spec): pass diff --git a/nfvbench/fluentd.py b/nfvbench/fluentd.py index 628b968..ad0ea34 100644 --- a/nfvbench/fluentd.py +++ b/nfvbench/fluentd.py @@ -35,7 +35,7 @@ class FluentLogHandler(logging.Handler): logging.Handler.__init__(self) self.log_senders = [] self.result_senders = [] - self.runlogdate = 0 + self.runlogdate = "1970-01-01T00:00:00.000000+0000" self.formatter = logging.Formatter('%(message)s') for fluentd_config in fluentd_configs: if fluentd_config.logging_tag: @@ -65,8 +65,9 @@ class FluentLogHandler(logging.Handler): "message": self.formatter.format(record), "@timestamp": self.__get_timestamp() } - # if runlogdate is 0, it's a log from server (not an nfvbench run) so do not send runlogdate - if self.runlogdate != 0: + # if runlogdate is Jan 1st 1970, it's a log from server (not an nfvbench run) + # so do not send runlogdate + if self.runlogdate != "1970-01-01T00:00:00.000000+0000": data["runlogdate"] = self.runlogdate self.__update_stats(record.levelno) @@ -103,9 +104,9 @@ class FluentLogHandler(logging.Handler): "numwarnings": self.__warning_counter, "@timestamp": self.__get_timestamp() } - # if runlogdate is 0, it's a log from server (not an nfvbench run) + # if runlogdate is Jan 1st 1970, it's a log from server (not an nfvbench run) # so don't send runlogdate - if self.runlogdate != 0: + if self.runlogdate != "1970-01-01T00:00:00.000000+0000": data["runlogdate"] = self.runlogdate for log_sender in self.log_senders: log_sender.emit(None, data) diff --git a/nfvbench/network.py b/nfvbench/network.py index e097c2b..6c02f04 100644 --- a/nfvbench/network.py +++ b/nfvbench/network.py @@ -15,8 +15,10 @@ class Interface(object): + """A class to hold the RX and TX counters for a virtual or physical interface.""" def __init__(self, name, device, tx_packets, rx_packets): + """Create a new interface instance.""" self.name = name self.device = device self.packets = { @@ -25,38 +27,65 @@ class Interface(object): } def set_packets(self, tx, rx): + """Set tx and rx counters for this interface.""" self.packets = { 'tx': tx, 'rx': rx } def set_packets_diff(self, tx, rx): + """Subtract current counters from new set of counters and update with results.""" self.packets = { 'tx': tx - self.packets['tx'], 'rx': rx - self.packets['rx'], } def is_no_op(self): + """Check if this interface is a no-opn interface.""" return self.name is None def get_packet_count(self, traffic_type): + """Get packet count for given direction.""" return self.packets.get(traffic_type, 0) @staticmethod def no_op(): + """Return an interface that doe snot pass any traffic.""" return Interface(None, None, 0, 0) class Network(object): + """This class holds all interfaces that make up a logical neutron network. + + A loopback packet path has exactly 2 networks. + The first interface is always one of the 2 traffic gen interface. + Subsequent interfaces are sorted along the path from the TG to the loopback point + which could be interfaces in a switch, a vswitch or a VM. + """ def __init__(self, interfaces=None, reverse=False): + """Create a network with initial interface list and direction. + + :param interfaces: initial interface list + :param reverse: specifies the order of interfaces returned by get_interfaces + """ if interfaces is None: interfaces = [] self.interfaces = interfaces self.reverse = reverse def add_interface(self, interface): + """Add one more interface to this network. + + Order if important as interfaces must be added from traffic generator ports towards then + looping back device. + """ self.interfaces.append(interface) def get_interfaces(self): + """Get interfaces associated to this network. + + Returned interface list is ordered from traffic generator port towards looping device if + reverse is false. Else returms the list in the reverse order. + """ return self.interfaces[::-1] if self.reverse else self.interfaces diff --git a/nfvbench/nfvbench.py b/nfvbench/nfvbench.py index 7d2e037..e0b5786 100644 --- a/nfvbench/nfvbench.py +++ b/nfvbench/nfvbench.py @@ -30,6 +30,7 @@ from pkg_resources import resource_string from __init__ import __version__ from chain_runner import ChainRunner +from cleanup import Cleaner from config import config_load from config import config_loads import credentials as credentials @@ -49,6 +50,7 @@ fluent_logger = None class NFVBench(object): """Main class of NFV benchmarking tool.""" + STATUS_OK = 'OK' STATUS_ERROR = 'ERROR' @@ -68,7 +70,7 @@ class NFVBench(object): sys.stdout.flush() def setup(self): - self.specs.set_run_spec(self.config_plugin.get_run_spec(self.specs.openstack)) + self.specs.set_run_spec(self.config_plugin.get_run_spec(self.config, self.specs.openstack)) self.chain_runner = ChainRunner(self.config, self.clients, self.cred, @@ -91,7 +93,19 @@ class NFVBench(object): try: self.update_config(opts) self.setup() - + new_frame_sizes = [] + min_packet_size = "68" if self.config.vlan_tagging else "64" + for frame_size in self.config.frame_sizes: + try: + if int(frame_size) < int(min_packet_size): + new_frame_sizes.append(min_packet_size) + LOG.info("Adjusting frame size %s Bytes to minimum size %s Bytes due to " + + "traffic generator restriction", frame_size, min_packet_size) + else: + new_frame_sizes.append(frame_size) + except ValueError: + new_frame_sizes.append(frame_size) + self.config.actual_frame_sizes = tuple(new_frame_sizes) result = { "date": datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'), "nfvbench_version": __version__, @@ -129,7 +143,7 @@ class NFVBench(object): } def prepare_summary(self, result): - """Prepares summary of the result to print and send it to logger (eg: fluentd)""" + """Prepare summary of the result to print and send it to logger (eg: fluentd).""" global fluent_logger summary = NFVBenchSummarizer(result, fluent_logger) LOG.info(str(summary)) @@ -163,6 +177,15 @@ class NFVBench(object): self.config.duration_sec = float(self.config.duration_sec) self.config.interval_sec = float(self.config.interval_sec) + self.config.pause_sec = float(self.config.pause_sec) + + # Get traffic generator profile config + if not self.config.generator_profile: + self.config.generator_profile = self.config.traffic_generator.default_profile + + generator_factory = TrafficGeneratorFactory(self.config) + self.config.generator_config = \ + generator_factory.get_generator_config(self.config.generator_profile) # Check length of mac_addrs_left/right for serivce_chain EXT with no_arp if self.config.service_chain == ChainType.EXT and self.config.no_arp: @@ -183,14 +206,6 @@ class NFVBench(object): b=len(self.config.generator_config.mac_addrs_right), c=self.config.service_chain_count)) - # Get traffic generator profile config - if not self.config.generator_profile: - self.config.generator_profile = self.config.traffic_generator.default_profile - - generator_factory = TrafficGeneratorFactory(self.config) - self.config.generator_config = \ - generator_factory.get_generator_config(self.config.generator_profile) - if not any(self.config.generator_config.pcis): raise Exception("PCI addresses configuration for selected traffic generator profile " "({tg_profile}) are missing. Please specify them in configuration file." @@ -211,12 +226,12 @@ class NFVBench(object): if self.config.openrc_file: self.config.openrc_file = os.path.expanduser(self.config.openrc_file) - self.config.ndr_run = (not self.config.no_traffic - and 'ndr' in self.config.rate.strip().lower().split('_')) - self.config.pdr_run = (not self.config.no_traffic - and 'pdr' in self.config.rate.strip().lower().split('_')) - self.config.single_run = (not self.config.no_traffic - and not (self.config.ndr_run or self.config.pdr_run)) + self.config.ndr_run = (not self.config.no_traffic and + 'ndr' in self.config.rate.strip().lower().split('_')) + self.config.pdr_run = (not self.config.no_traffic and + 'pdr' in self.config.rate.strip().lower().split('_')) + self.config.single_run = (not self.config.no_traffic and + not (self.config.ndr_run or self.config.pdr_run)) if self.config.vlans and len(self.config.vlans) != 2: raise Exception('Number of configured VLAN IDs for VLAN tagging must be exactly 2.') @@ -240,6 +255,11 @@ class NFVBench(object): def parse_opts_from_cli(): parser = argparse.ArgumentParser() + parser.add_argument('--status', dest='status', + action='store_true', + default=None, + help='Provide NFVbench status') + parser.add_argument('-c', '--config', dest='config', action='store', help='Override default values with a config file or ' @@ -354,6 +374,16 @@ def parse_opts_from_cli(): action='store_true', help='no cleanup after run') + parser.add_argument('--cleanup', dest='cleanup', + default=None, + action='store_true', + help='Cleanup NFVbench resources (prompt to confirm)') + + parser.add_argument('--force-cleanup', dest='force_cleanup', + default=None, + action='store_true', + help='Cleanup NFVbench resources (do not prompt)') + parser.add_argument('--json', dest='json', action='store', help='store results in json format file', @@ -405,6 +435,11 @@ def parse_opts_from_cli(): action='store', help='Custom label for performance records') + parser.add_argument('--l2-loopback', '--l2loopback', dest='l2_loopback', + action='store', + metavar='<vlan>', + help='Port to port or port to switch to port L2 loopback with VLAN id') + opts, unknown_opts = parser.parse_known_args() return opts, unknown_opts @@ -417,8 +452,7 @@ def load_default_config(): def override_custom_traffic(config, frame_sizes, unidir): - """Override the traffic profiles with a custom one - """ + """Override the traffic profiles with a custom one.""" if frame_sizes is not None: traffic_profile_name = "custom_traffic_profile" config.traffic_profile = [ @@ -445,6 +479,23 @@ def check_physnet(name, netattrs): raise Exception("SRIOV requires segmentation_id to be specified for the {n} network" .format(n=name)) +def status_cleanup(config, cleanup, force_cleanup): + LOG.info('Version: %s', pbr.version.VersionInfo('nfvbench').version_string_with_vcs()) + # check if another run is pending + ret_code = 0 + try: + with utils.RunLock(): + LOG.info('Status: idle') + except Exception: + LOG.info('Status: busy (run pending)') + ret_code = 1 + # check nfvbench resources + if config.openrc_file and config.service_chain != ChainType.EXT: + cleaner = Cleaner(config) + count = cleaner.show_resources() + if count and (cleanup or force_cleanup): + cleaner.clean(not force_cleanup) + sys.exit(ret_code) def main(): global fluent_logger @@ -467,14 +518,6 @@ def main(): opts, unknown_opts = parse_opts_from_cli() log.set_level(debug=opts.debug) - # setup the fluent logger as soon as possible right after the config plugin is called, - # if there is any logging or result tag is set then initialize the fluent logger - for fluentd in config.fluentd: - if fluentd.logging_tag or fluentd.result_tag: - fluent_logger = FluentLogHandler(config.fluentd) - LOG.addHandler(fluent_logger) - break - if opts.version: print pbr.version.VersionInfo('nfvbench').version_string_with_vcs() sys.exit(0) @@ -506,6 +549,14 @@ def main(): LOG.info('Loading configuration string: %s', opts.config) config = config_loads(opts.config, config, whitelist_keys) + # setup the fluent logger as soon as possible right after the config plugin is called, + # if there is any logging or result tag is set then initialize the fluent logger + for fluentd in config.fluentd: + if fluentd.logging_tag or fluentd.result_tag: + fluent_logger = FluentLogHandler(config.fluentd) + LOG.addHandler(fluent_logger) + break + # traffic profile override options override_custom_traffic(config, opts.frame_sizes, opts.unidir) @@ -524,8 +575,22 @@ def main(): if opts.no_int_config: config.no_int_config = opts.no_int_config + # port to port loopback (direct or through switch) + if opts.l2_loopback: + config.l2_loopback = True + if config.service_chain != ChainType.EXT: + LOG.info('Changing service chain type to EXT') + config.service_chain = ChainType.EXT + if not config.no_arp: + LOG.info('Disabling ARP') + config.no_arp = True + config.vlans = [int(opts.l2_loopback), int(opts.l2_loopback)] + # disable any form of interface config since we loop at the switch level + config.no_int_config = True + LOG.info('Running L2 loopback: using EXT chain/no ARP') + if opts.use_sriov_middle_net: - if (not config.sriov) or (not config.service_chain == ChainType.PVVP): + if (not config.sriov) or (config.service_chain != ChainType.PVVP): raise Exception("--use-sriov-middle-net is only valid for PVVP with SRIOV") config.use_sriov_middle_net = True @@ -554,6 +619,9 @@ def main(): # in a copy of the dict (config plugin still holds the original dict) config_plugin.set_config(config) + if opts.status or opts.cleanup or opts.force_cleanup: + status_cleanup(config, opts.cleanup, opts.force_cleanup) + # add file log if requested if config.log_file: log.add_file_logger(config.log_file) diff --git a/nfvbench/nfvbenchd.py b/nfvbench/nfvbenchd.py index 76906c5..fa781af 100644 --- a/nfvbench/nfvbenchd.py +++ b/nfvbench/nfvbenchd.py @@ -226,7 +226,7 @@ class WebSocketIoServer(object): # print 'main thread waiting for requests...' config = Ctx.dequeue() # print 'main thread processing request...' - print config + # print config try: # remove unfilled values as we do not want them to override default values with None config = {k: v for k, v in config.items() if v is not None} @@ -243,8 +243,15 @@ class WebSocketIoServer(object): else: # this might overwrite a previously unfetched result Ctx.set_result(results) - summary = NFVBenchSummarizer(results['result'], self.fluent_logger) - LOG.info(str(summary)) + try: + summary = NFVBenchSummarizer(results['result'], self.fluent_logger) + LOG.info(str(summary)) + except KeyError: + # in case of error, 'result' might be missing + if 'error_message' in results: + LOG.error(results['error_message']) + else: + LOG.error('REST request completed without results or error message') Ctx.release() if self.fluent_logger: self.fluent_logger.send_run_summary(True) diff --git a/nfvbench/service_chain.py b/nfvbench/service_chain.py index 216cc92..7ec1511 100644 --- a/nfvbench/service_chain.py +++ b/nfvbench/service_chain.py @@ -57,14 +57,14 @@ class ServiceChain(object): for vlan, device in zip(vlans, self.config.generator_config.devices): self.stats_manager.set_vlan_tag(device, vlan) - def __get_result_per_frame_size(self, frame_size, bidirectional): + def __get_result_per_frame_size(self, frame_size, actual_frame_size, bidirectional): start_time = time.time() traffic_result = { frame_size: {} } result = {} if not self.config.no_traffic: - self.clients['traffic'].set_traffic(frame_size, bidirectional) + self.clients['traffic'].set_traffic(actual_frame_size, bidirectional) if self.config.single_run: result = self.stats_manager.run() @@ -73,6 +73,9 @@ class ServiceChain(object): for dr in ['pdr', 'ndr']: if dr in results: + if frame_size != actual_frame_size: + results[dr]['l2frame_size'] = frame_size + results[dr]['actual_l2frame_size'] = actual_frame_size traffic_result[frame_size][dr] = results[dr] if 'warning' in results[dr]['stats'] and results[dr]['stats']['warning']: traffic_result['warning'] = results[dr]['stats']['warning'] @@ -83,6 +86,8 @@ class ServiceChain(object): result['run_config'] = self.clients['traffic'].get_run_config(result) required = result['run_config']['direction-total']['orig']['rate_pps'] actual = result['stats']['total_tx_rate'] + if frame_size != actual_frame_size: + result['actual_l2frame_size'] = actual_frame_size warning = self.clients['traffic'].compare_tx_rates(required, actual) if warning is not None: result['run_config']['warning'] = warning @@ -92,8 +97,10 @@ class ServiceChain(object): def __get_chain_result(self): result = OrderedDict() - for fs in self.config.frame_sizes: - result.update(self.__get_result_per_frame_size(fs, self.config.traffic.bidirectional)) + for fs, actual_fs in zip(self.config.frame_sizes, self.config.actual_frame_sizes): + result.update(self.__get_result_per_frame_size(fs, + actual_fs, + self.config.traffic.bidirectional)) chain_result = { 'flow_count': self.config.flow_count, diff --git a/nfvbench/summarizer.py b/nfvbench/summarizer.py index 0ff9c48..b27ed6f 100644 --- a/nfvbench/summarizer.py +++ b/nfvbench/summarizer.py @@ -312,11 +312,17 @@ class NFVBenchSummarizer(Summarizer): if 'warning' in entry: continue self.__chain_analysis_summarize(*entry) - self.__record_send() + self.__record_send() def __chain_analysis_summarize(self, frame_size, analysis): self._put() self._put('L2 frame size:', frame_size) + if 'actual_l2frame_size' in analysis: + self._put('Actual l2 frame size:', analysis['actual_l2frame_size']) + elif self.config['ndr_run'] and 'actual_l2frame_size' in analysis['ndr']: + self._put('Actual l2 frame size:', analysis['ndr']['actual_l2frame_size']) + elif self.config['pdr_run'] and 'actual_l2frame_size' in analysis['pdr']: + self._put('Actual l2 frame size:', analysis['pdr']['actual_l2frame_size']) if 'analysis_duration_sec' in analysis: self._put('Chain analysis duration:', Formatter.float(3)(analysis['analysis_duration_sec']), 'seconds') diff --git a/nfvbench/traffic_client.py b/nfvbench/traffic_client.py index 57141be..ef68fe5 100755 --- a/nfvbench/traffic_client.py +++ b/nfvbench/traffic_client.py @@ -35,10 +35,14 @@ from utils import cast_integer class TrafficClientException(Exception): + """Generic traffic client exception.""" + pass class TrafficRunner(object): + """Serialize various steps required to run traffic.""" + def __init__(self, client, duration_sec, interval_sec=0): self.client = client self.start_time = None @@ -89,6 +93,8 @@ class TrafficRunner(object): class IpBlock(object): + """Manage a block of IP addresses.""" + def __init__(self, base_ip, step_ip, count_ip): self.base_ip_int = Device.ip_to_int(base_ip) self.step = Device.ip_to_int(step_ip) @@ -96,15 +102,13 @@ class IpBlock(object): self.next_free = 0 def get_ip(self, index=0): - '''Return the IP address at given index - ''' + """Return the IP address at given index.""" if index < 0 or index >= self.max_available: raise IndexError('Index out of bounds') return Device.int_to_ip(self.base_ip_int + index * self.step) def reserve_ip_range(self, count): - '''Reserve a range of count consecutive IP addresses spaced by step - ''' + """Reserve a range of count consecutive IP addresses spaced by step.""" if self.next_free + count > self.max_available: raise IndexError('No more IP addresses next free=%d max_available=%d requested=%d' % (self.next_free, @@ -120,6 +124,8 @@ class IpBlock(object): class Device(object): + """Represent a port device and all information associated to it.""" + def __init__(self, port, pci, switch_port=None, vtep_vlan=None, ip=None, tg_gateway_ip=None, gateway_ip=None, ip_addrs_step=None, tg_gateway_ip_addrs_step=None, gateway_ip_addrs_step=None, udp_src_port=None, udp_dst_port=None, @@ -158,6 +164,7 @@ class Device(object): if mac is None: raise TrafficClientException('Trying to set traffic generator MAC address as None') self.mac = mac + LOG.info("Port %d: src MAC %s", self.port, self.mac) def set_destination(self, dst): self.dst = dst @@ -169,10 +176,10 @@ class Device(object): if self.vlan_tagging and vlan_tag is None: raise TrafficClientException('Trying to set VLAN tag as None') self.vlan_tag = vlan_tag + LOG.info("Port %d: VLAN %d", self.port, self.vlan_tag) def get_gw_ip(self, chain_index): - '''Retrieve the IP address assigned for the gateway of a given chain - ''' + """Retrieve the IP address assigned for the gateway of a given chain.""" return self.gw_ip_block.get_ip(chain_index) def get_stream_configs(self, service_chain): @@ -222,8 +229,7 @@ class Device(object): return configs def ip_range_overlaps(self): - '''Check if this device ip range is overlapping with the dst device ip range - ''' + """Check if this device ip range is overlapping with the dst device ip range.""" src_base_ip = Device.ip_to_int(self.ip) dst_base_ip = Device.ip_to_int(self.dst.ip) src_last_ip = src_base_ip + self.flow_count - 1 @@ -367,6 +373,8 @@ class RunningTrafficProfile(object): class TrafficGeneratorFactory(object): + """Factory class to generate a traffic generator.""" + def __init__(self, config): self.config = config @@ -406,6 +414,8 @@ class TrafficGeneratorFactory(object): class TrafficClient(object): + """Traffic generator client.""" + PORTS = [0, 1] def __init__(self, config, notifier=None, skip_sleep=False): @@ -449,55 +459,56 @@ class TrafficClient(object): return self.gen.get_version() def ensure_end_to_end(self): - """ - Ensure traffic generator receives packets it has transmitted. + """Ensure traffic generator receives packets it has transmitted. + This ensures end to end connectivity and also waits until VMs are ready to forward packets. - At this point all VMs are in active state, but forwarding does not have to work. - Small amount of traffic is sent to every chain. Then total of sent and received packets - is compared. If ratio between received and transmitted packets is higher than (N-1)/N, - N being number of chains, traffic flows through every chain and real measurements can be - performed. + VMs that are started and in active state may not pass traffic yet. It is imperative to make + sure that all VMs are passing traffic in both directions before starting any benchmarking. + To verify this, we need to send at a low frequency bi-directional packets and make sure + that we receive all packets back from all VMs. The number of flows is equal to 2 times + the number of chains (1 per direction) and we need to make sure we receive packets coming + from exactly 2 x chain count different source MAC addresses. Example: PVP chain (1 VM per chain) N = 10 (number of chains) - threshold = (N-1)/N = 9/10 = 0.9 (acceptable ratio ensuring working conditions) - if total_received/total_sent > 0.9, traffic is flowing to more than 9 VMs meaning - all 10 VMs are in operational state. + Flow count = 20 (number of flows) + If the number of unique source MAC addresses from received packets is 20 then + all 10 VMs 10 VMs are in operational state. """ LOG.info('Starting traffic generator to ensure end-to-end connectivity') - rate_pps = {'rate_pps': str(self.config.service_chain_count * 100)} + rate_pps = {'rate_pps': str(self.config.service_chain_count * 1)} self.gen.create_traffic('64', [rate_pps, rate_pps], bidirectional=True, latency=False) # ensures enough traffic is coming back - threshold = (self.config.service_chain_count - 1) / float(self.config.service_chain_count) retry_count = (self.config.check_traffic_time_sec + self.config.generic_poll_sec - 1) / self.config.generic_poll_sec + mac_addresses = set() + ln = 0 + # in case of l2-loopback, we will only have 2 unique src MAC regardless of the + # number of chains configured because there are no VM involved + # otherwise, we expect to see packets coming from 2 unique MAC per chain + unique_src_mac_count = 2 if self.config.l2_loopback else self.config.service_chain_count * 2 for it in xrange(retry_count): self.gen.clear_stats() self.gen.start_traffic() + self.gen.start_capture() LOG.info('Waiting for packets to be received back... (%d / %d)', it + 1, retry_count) if not self.skip_sleep: time.sleep(self.config.generic_poll_sec) self.gen.stop_traffic() - stats = self.gen.get_stats() - - # compute total sent and received traffic on both ports - total_rx = 0 - total_tx = 0 - for port in self.PORTS: - total_rx += float(stats[port]['rx'].get('total_pkts', 0)) - total_tx += float(stats[port]['tx'].get('total_pkts', 0)) - - # how much of traffic came back - ratio = total_rx / total_tx if total_tx else 0 - - if ratio > threshold: - self.gen.clear_stats() - self.gen.clear_streamblock() - LOG.info('End-to-end connectivity ensured') - return + self.gen.fetch_capture_packets() + self.gen.stop_capture() + + for packet in self.gen.packet_list: + mac_addresses.add(packet['binary'][6:12]) + if ln != len(mac_addresses): + ln = len(mac_addresses) + LOG.info('Received unique source MAC %d / %d', ln, unique_src_mac_count) + if len(mac_addresses) == unique_src_mac_count: + LOG.info('End-to-end connectivity ensured') + return if not self.skip_sleep: time.sleep(self.config.generic_poll_sec) @@ -518,6 +529,10 @@ class TrafficClient(object): unidir_reverse_pps = int(self.config.unidir_reverse_traffic_pps) if unidir_reverse_pps > 0: self.run_config['rates'].append({'rate_pps': str(unidir_reverse_pps)}) + # Fix for [NFVBENCH-67], convert the rate string to PPS + for idx, rate in enumerate(self.run_config['rates']): + if 'rate_pps' not in rate: + self.run_config['rates'][idx] = {'rate_pps': self.__convert_rates(rate)['rate_pps']} self.gen.clear_streamblock() self.gen.create_traffic(frame_size, self.run_config['rates'], bidirectional, latency=True) @@ -660,7 +675,7 @@ class TrafficClient(object): results[tag]['timestamp_sec'] = time.time() def __range_search(self, left, right, targets, results): - '''Perform a binary search for a list of targets inside a [left..right] range or rate + """Perform a binary search for a list of targets inside a [left..right] range or rate. left the left side of the range to search as a % the line rate (100 = 100% line rate) indicating the rate to send on each interface @@ -669,7 +684,7 @@ class TrafficClient(object): targets a dict of drop rates to search (0.1 = 0.1%), indexed by the DR name or "tag" ('ndr', 'pdr') results a dict to store results - ''' + """ if not targets: return LOG.info('Range search [%s .. %s] targets: %s', left, right, targets) @@ -739,6 +754,7 @@ class TrafficClient(object): time_elapsed_ratio = self.runner.time_elapsed() / self.run_config['duration_sec'] if time_elapsed_ratio >= 1: self.cancel_traffic() + time.sleep(self.config.pause_sec) self.interval_collector.reset() # get stats from the run @@ -785,13 +801,11 @@ class TrafficClient(object): def cancel_traffic(self): self.runner.stop() - def get_interface(self, port_index): + def get_interface(self, port_index, stats): port = self.gen.port_handle[port_index] tx, rx = 0, 0 - if not self.config.no_traffic: - stats = self.get_stats() - if port in stats: - tx, rx = int(stats[port]['tx']['total_pkts']), int(stats[port]['rx']['total_pkts']) + if stats and port in stats: + tx, rx = int(stats[port]['tx']['total_pkts']), int(stats[port]['rx']['total_pkts']) return Interface('traffic-generator', self.tool.lower(), tx, rx) def get_traffic_config(self): @@ -820,11 +834,13 @@ class TrafficClient(object): return config def get_run_config(self, results): - """Returns configuration which was used for the last run.""" + """Return configuration which was used for the last run.""" r = {} + # because we want each direction to have the far end RX rates, + # use the far end index (1-idx) to retrieve the RX rates for idx, key in enumerate(["direction-forward", "direction-reverse"]): tx_rate = results["stats"][idx]["tx"]["total_pkts"] / self.config.duration_sec - rx_rate = results["stats"][idx]["rx"]["total_pkts"] / self.config.duration_sec + rx_rate = results["stats"][1 - idx]["rx"]["total_pkts"] / self.config.duration_sec r[key] = { "orig": self.__convert_rates(self.run_config['rates'][idx]), "tx": self.__convert_rates({'rate_pps': tx_rate}), @@ -835,7 +851,6 @@ class TrafficClient(object): for direction in ['orig', 'tx', 'rx']: total[direction] = {} for unit in ['rate_percent', 'rate_bps', 'rate_pps']: - total[direction][unit] = sum([float(x[direction][unit]) for x in r.values()]) r['direction-total'] = total diff --git a/nfvbench/traffic_gen/dummy.py b/nfvbench/traffic_gen/dummy.py index 6f57f4d..788a53f 100644 --- a/nfvbench/traffic_gen/dummy.py +++ b/nfvbench/traffic_gen/dummy.py @@ -31,6 +31,11 @@ class DummyTG(AbstractTrafficGenerator): self.duration_sec = self.config.duration_sec self.intf_speed = config.generator_config.intf_speed self.set_response_curve() + self.packet_list = [{ + "binary": "01234567890123456789" + }, { + "binary": "98765432109876543210" + }] def get_version(self): return "0.1" @@ -164,9 +169,18 @@ class DummyTG(AbstractTrafficGenerator): def start_traffic(self): pass + def fetch_capture_packets(self): + pass + def stop_traffic(self): pass + def start_capture(self): + pass + + def stop_capture(self): + pass + def cleanup(self): pass diff --git a/nfvbench/traffic_gen/traffic_base.py b/nfvbench/traffic_gen/traffic_base.py index 817ecc8..81537b3 100644 --- a/nfvbench/traffic_gen/traffic_base.py +++ b/nfvbench/traffic_gen/traffic_base.py @@ -23,19 +23,12 @@ class TrafficGeneratorException(Exception): class AbstractTrafficGenerator(object): - # src_mac (6) + dst_mac (6) + mac_type (2) + frame_check (4) = 18 - l2_header_size = 18 - - imix_l2_sizes = [64, 594, 1518] - imix_l3_sizes = [size - l2_header_size for size in imix_l2_sizes] - imix_ratios = [7, 4, 1] - imix_avg_l2_size = sum( - [1.0 * imix[0] * imix[1] for imix in zip(imix_l2_sizes, imix_ratios)]) / sum(imix_ratios) - - traffic_utils.imix_avg_l2_size = imix_avg_l2_size - def __init__(self, config): self.config = config + self.imix_l2_sizes = [64, 594, 1518] + self.imix_ratios = [7, 4, 1] + self.imix_avg_l2_size = 0 + self.adjust_imix_min_size(64) @abc.abstractmethod def get_version(self): @@ -58,7 +51,7 @@ class AbstractTrafficGenerator(object): return None @abc.abstractmethod - def create_traffic(self): + def create_traffic(self, l2frame_size, rates, bidirectional, latency=True): # Must be implemented by sub classes return None @@ -96,3 +89,11 @@ class AbstractTrafficGenerator(object): def cleanup(self): # Must be implemented by sub classes return None + + def adjust_imix_min_size(self, min_size): + # assume the min size is always the first entry + self.imix_l2_sizes[0] = min_size + self.imix_avg_l2_size = sum( + [1.0 * imix[0] * imix[1] for imix in zip(self.imix_l2_sizes, self.imix_ratios)]) / sum( + self.imix_ratios) + traffic_utils.imix_avg_l2_size = self.imix_avg_l2_size diff --git a/nfvbench/traffic_gen/trex.py b/nfvbench/traffic_gen/trex.py index 23faebc..cabf1cb 100644 --- a/nfvbench/traffic_gen/trex.py +++ b/nfvbench/traffic_gen/trex.py @@ -48,6 +48,8 @@ from trex_stl_lib.api import STLVmFlowVarRepetableRandom from trex_stl_lib.api import STLVmWrFlowVar from trex_stl_lib.api import UDP from trex_stl_lib.services.trex_stl_service_arp import STLServiceARP + + # pylint: enable=import-error @@ -64,31 +66,42 @@ class TRex(AbstractTrafficGenerator): self.streamblock = defaultdict(list) self.rates = [] self.arps = {} + self.capture_id = None + self.packet_list = [] def get_version(self): return self.client.get_server_version() def extract_stats(self, in_stats): + """Extract stats from dict returned by Trex API. + + :param in_stats: dict as returned by TRex api + """ utils.nan_replace(in_stats) LOG.debug(in_stats) result = {} + # port_handles should have only 2 elements: [0, 1] + # so (1 - ph) will be the index for the far end port for ph in self.port_handle: - stats = self.__combine_stats(in_stats, ph) + stats = in_stats[ph] + far_end_stats = in_stats[1 - ph] result[ph] = { 'tx': { - 'total_pkts': cast_integer(stats['tx_pkts']['total']), - 'total_pkt_bytes': cast_integer(stats['tx_bytes']['total']), - 'pkt_rate': cast_integer(stats['tx_pps']['total']), - 'pkt_bit_rate': cast_integer(stats['tx_bps']['total']) + 'total_pkts': cast_integer(stats['opackets']), + 'total_pkt_bytes': cast_integer(stats['obytes']), + 'pkt_rate': cast_integer(stats['tx_pps']), + 'pkt_bit_rate': cast_integer(stats['tx_bps']) }, 'rx': { - 'total_pkts': cast_integer(stats['rx_pkts']['total']), - 'total_pkt_bytes': cast_integer(stats['rx_bytes']['total']), - 'pkt_rate': cast_integer(stats['rx_pps']['total']), - 'pkt_bit_rate': cast_integer(stats['rx_bps']['total']), + 'total_pkts': cast_integer(stats['ipackets']), + 'total_pkt_bytes': cast_integer(stats['ibytes']), + 'pkt_rate': cast_integer(stats['rx_pps']), + 'pkt_bit_rate': cast_integer(stats['rx_bps']), + # how many pkts were dropped in RX direction + # need to take the tx counter on the far end port 'dropped_pkts': cast_integer( - stats['tx_pkts']['total'] - stats['rx_pkts']['total']) + far_end_stats['opackets'] - stats['ipackets']) } } @@ -103,20 +116,6 @@ class TRex(AbstractTrafficGenerator): result["total_tx_rate"] = cast_integer(total_tx_pkts / self.config.duration_sec) return result - def __combine_stats(self, in_stats, port_handle): - """Traverses TRex result dictionary and combines stream stats. Used for combining latency - and regular streams together. - """ - result = defaultdict(lambda: defaultdict(float)) - - for pg_id in [self.stream_ids[port_handle]] + self.latencies[port_handle]: - record = in_stats['flow_stats'][pg_id] - for stat_type, stat_type_values in record.iteritems(): - for ph, value in stat_type_values.iteritems(): - result[stat_type][ph] += value - - return result - def __combine_latencies(self, in_stats, port_handle): """Traverses TRex result dictionary and combines chosen latency stats.""" if not self.latencies[port_handle]: @@ -136,14 +135,16 @@ class TRex(AbstractTrafficGenerator): return result def create_pkt(self, stream_cfg, l2frame_size): - # 46 = 14 (Ethernet II) + 4 (CRC Checksum) + 20 (IPv4) + 8 (UDP) - payload = 'x' * (max(64, int(l2frame_size)) - 46) pkt_base = Ether(src=stream_cfg['mac_src'], dst=stream_cfg['mac_dst']) - if stream_cfg['vlan_tag'] is not None: + # 50 = 14 (Ethernet II) + 4 (Vlan tag) + 4 (CRC Checksum) + 20 (IPv4) + 8 (UDP) pkt_base /= Dot1Q(vlan=stream_cfg['vlan_tag']) - + l2payload_size = int(l2frame_size) - 50 + else: + # 46 = 14 (Ethernet II) + 4 (CRC Checksum) + 20 (IPv4) + 8 (UDP) + l2payload_size = int(l2frame_size) - 46 + payload = 'x' * l2payload_size udp_args = {} if stream_cfg['udp_src_port']: udp_args['sport'] = int(stream_cfg['udp_src_port']) @@ -198,6 +199,8 @@ class TRex(AbstractTrafficGenerator): idx_lat = None streams = [] if l2frame == 'IMIX': + min_size = 64 if stream_cfg['vlan_tag'] is None else 68 + self.adjust_imix_min_size(min_size) for t, (ratio, l2_frame_size) in enumerate(zip(self.imix_ratios, self.imix_l2_sizes)): pkt = self.create_pkt(stream_cfg, l2_frame_size) streams.append(STLStream(packet=pkt, @@ -208,6 +211,7 @@ class TRex(AbstractTrafficGenerator): if latency: idx_lat = self.id.next() + pkt = self.create_pkt(stream_cfg, self.imix_avg_l2_size) sl = STLStream(packet=pkt, isg=isg, flow_stats=STLFlowLatencyStats(pg_id=idx_lat), @@ -247,7 +251,7 @@ class TRex(AbstractTrafficGenerator): break except Exception as ex: if it == (self.config.generic_retry_count - 1): - raise ex + raise LOG.info("Retrying connection to TRex (%s)...", ex.message) def connect(self): @@ -318,7 +322,7 @@ class TRex(AbstractTrafficGenerator): def __start_server(self): server = TRexTrafficServer() - server.run_server(self.config.generator_config) + server.run_server(self.config.generator_config, self.config.vlan_tagging) def resolve_arp(self): self.client.set_service_mode(ports=self.port_handle) @@ -354,7 +358,7 @@ class TRex(AbstractTrafficGenerator): else: failed = [arp.get_record().dst_ip for arp in arps if arp.get_record().dst_mac is None] - LOG.info('Retrying ARP for: %d (%d / %d)', + LOG.info('Retrying ARP for: %s (%d / %d)', failed, attempt, self.config.generic_retry_count) time.sleep(self.config.generic_poll_sec) @@ -433,7 +437,7 @@ class TRex(AbstractTrafficGenerator): LOG.info('Cleared all existing streams.') def get_stats(self): - stats = self.client.get_pgid_stats() + stats = self.client.get_stats() return self.extract_stats(stats) def get_macs(self): @@ -450,6 +454,31 @@ class TRex(AbstractTrafficGenerator): def stop_traffic(self): self.client.stop(ports=self.port_handle) + def start_capture(self): + """Capture all packets on both ports that are unicast to us.""" + if self.capture_id: + self.stop_capture() + # Need to filter out unwanted packets so we do not end up counting + # src MACs of frames that are not unicast to us + src_mac_list = self.get_macs() + bpf_filter = "ether dst %s or ether dst %s" % (src_mac_list[0], src_mac_list[1]) + # ports must be set in service in order to enable capture + self.client.set_service_mode(ports=self.port_handle) + self.capture_id = self.client.start_capture(rx_ports=self.port_handle, + bpf_filter=bpf_filter) + + def fetch_capture_packets(self): + if self.capture_id: + self.packet_list = [] + self.client.fetch_capture_packets(capture_id=self.capture_id['id'], + output=self.packet_list) + + def stop_capture(self): + if self.capture_id: + self.client.stop_capture(capture_id=self.capture_id['id']) + self.capture_id = None + self.client.set_service_mode(ports=self.port_handle, enabled=False) + def cleanup(self): if self.client: try: diff --git a/nfvbench/traffic_server.py b/nfvbench/traffic_server.py index fe9edd2..dcb83fb 100644 --- a/nfvbench/traffic_server.py +++ b/nfvbench/traffic_server.py @@ -34,7 +34,7 @@ class TRexTrafficServer(TrafficServer): assert len(contents) == 1 self.trex_dir = os.path.join(trex_base_dir, contents[0]) - def run_server(self, traffic_profile, filename='/etc/trex_cfg.yaml'): + def run_server(self, traffic_profile, vlan_tagging, filename='/etc/trex_cfg.yaml'): """ Runs TRex server for specified traffic profile. @@ -43,10 +43,12 @@ class TRexTrafficServer(TrafficServer): """ cfg = self.__save_config(traffic_profile, filename) cores = traffic_profile.cores - sw_mode = "--software" if traffic_profile.software_mode else "" + sw_mode = "--software" if traffic_profile.generator_config.software_mode else "" + vlan_opt = "--vlan" if vlan_tagging else "" subprocess.Popen(['nohup', '/bin/bash', '-c', './t-rex-64 -i -c {} --iom 0 --no-scapy-server --close-at-end {} ' - '--vlan --cfg {} &> /tmp/trex.log & disown'.format(cores, sw_mode, cfg)], + '{} --cfg {} &> /tmp/trex.log & disown'.format(cores, sw_mode, + vlan_opt, cfg)], cwd=self.trex_dir) LOG.info('TRex server is running...') diff --git a/nfvbench/utils.py b/nfvbench/utils.py index 20dc588..ecbb55a 100644 --- a/nfvbench/utils.py +++ b/nfvbench/utils.py @@ -12,6 +12,7 @@ # License for the specific language governing permissions and limitations # under the License. +import glob from math import isnan import os import re @@ -91,11 +92,45 @@ def dict_to_json_dict(record): return json.loads(json.dumps(record, default=lambda obj: obj.to_json())) -def get_intel_pci(nic_ports): - """Returns the first two PCI addresses of sorted PCI list for Intel NIC (i40e, ixgbe)""" - hx = r'[0-9a-fA-F]' - regex = r'{hx}{{4}}:({hx}{{2}}:{hx}{{2}}\.{hx}{{1}}).*(drv={driver}|.*unused=.*{driver})' +def get_intel_pci(nic_slot=None, nic_ports=None): + """Returns two PCI address that will be used for NFVbench + + @param nic_slot: The physical PCIe slot number in motherboard + @param nic_ports: Array of two integers indicating the ports to use on the NIC + + When nic_slot and nic_ports are both supplied, the function will just return + the PCI addresses for them. The logic used is: + (1) Run "dmidecode -t slot" + (2) Grep for "SlotID:" with given nic_slot, and derive the bus address; + (3) Based on given nic_ports, generate the pci addresses based on above + base address; + + When either nic_slot or nic_ports is not supplied, the function will + traverse all Intel NICs which use i40e or ixgbe driver, sorted by PCI + address, and return first two available ports which are not bonded + (802.11ad). + """ + if nic_slot and nic_ports: + dmidecode = subprocess.check_output(['dmidecode', '-t', 'slot']) + regex = r"(?<=SlotID:%s).*?(....:..:..\..)" % nic_slot + match = re.search(regex, dmidecode, flags=re.DOTALL) + if not match: + return None + + pcis = [] + # On some servers, the "Bus Address" returned by dmidecode is not the + # base pci address of the NIC. So only keeping the bus part of the + # address for better compability. + bus = match.group(1)[:match.group(1).rindex(':') + 1] + "00." + for port in nic_ports: + pcis.append(bus + str(port)) + + return pcis + + hx = r'[0-9a-fA-F]' + regex = r'({hx}{{4}}:({hx}{{2}}:{hx}{{2}}\.{hx}{{1}})).*(drv={driver}|.*unused=.*{driver})' + pcis = [] try: trex_base_dir = '/opt/trex' contents = os.listdir(trex_base_dir) @@ -110,14 +145,28 @@ def get_intel_pci(nic_ports): for driver in ['i40e', 'ixgbe']: matches = re.findall(regex.format(hx=hx, driver=driver), devices) - if matches: - pcis = [x[0] for x in matches] - if len(pcis) < 2: - continue - pcis.sort() - return [pcis[port_index] for port_index in nic_ports] - - return [] + if not matches: + continue + + matches.sort() + for port in matches: + intf_name = glob.glob("/sys/bus/pci/devices/%s/net/*" % port[0]) + if not intf_name: + # Interface is not bind to kernel driver, so take it + pcis.append(port[1]) + else: + intf_name = intf_name[0][intf_name[0].rfind('/') + 1:] + process = subprocess.Popen(['ip', '-o', '-d', 'link', 'show', intf_name], + stdout=subprocess.PIPE, + stderr=subprocess.PIPE) + intf_info, _ = process.communicate() + if not re.search('team_slave|bond_slave', intf_info): + pcis.append(port[1]) + + if len(pcis) == 2: + break + + return pcis multiplier_map = { diff --git a/nfvbenchvm/dib/build-image.sh b/nfvbenchvm/dib/build-image.sh index 88cf2e5..d9ab20e 100755 --- a/nfvbenchvm/dib/build-image.sh +++ b/nfvbenchvm/dib/build-image.sh @@ -11,7 +11,7 @@ set -e gs_url=artifacts.opnfv.org/nfvbench/images # image version number -__version__=0.5 +__version__=0.6 image_name=nfvbenchvm_centos-$__version__ # if image exists skip building diff --git a/nfvbenchvm/dib/elements/nfvbenchvm/post-install.d/02-testpmd-script b/nfvbenchvm/dib/elements/nfvbenchvm/post-install.d/02-testpmd-script index 2136c3a..6f133d1 100755 --- a/nfvbenchvm/dib/elements/nfvbenchvm/post-install.d/02-testpmd-script +++ b/nfvbenchvm/dib/elements/nfvbenchvm/post-install.d/02-testpmd-script @@ -1,7 +1,7 @@ #!/bin/bash -DPDK=dpdk-17.08 -DPDK_UNTAR=dpdk-17.08 +DPDK=dpdk-18.02 +DPDK_UNTAR=dpdk-18.02 # pick up the kernel version for the target image kernel_version=`ls -t /lib/modules | awk 'NR==1 {print}'` diff --git a/nfvbenchvm/dib/elements/nfvbenchvm/static/etc/rc.d/rc.local b/nfvbenchvm/dib/elements/nfvbenchvm/static/etc/rc.d/rc.local index dddc787..caf3142 100644 --- a/nfvbenchvm/dib/elements/nfvbenchvm/static/etc/rc.d/rc.local +++ b/nfvbenchvm/dib/elements/nfvbenchvm/static/etc/rc.d/rc.local @@ -71,10 +71,8 @@ if [ "$FORWARDER" == "testpmd" ]; then -n 4 \ -- \ --burst=32 \ - --txd=2048 \ - --rxd=2048 \ - --txqflags=0xf00 \ - --disable-hw-vlan \ + --txd=256 \ + --rxd=1024 \ --eth-peer=0,$TG_MAC1 \ --eth-peer=1,$TG_MAC2 \ --forward-mode=mac \ @@ -40,7 +40,6 @@ packages = console_scripts = nfvbench = nfvbench.nfvbench:main nfvbench_client = client.nfvbench_client:main - nfvbench_cleanup = cleanup.nfvbench_cleanup:main [compile_catalog] directory = nfvbench/locale diff --git a/test/test_nfvbench.py b/test/test_nfvbench.py index 16784d8..4603ae1 100644 --- a/test/test_nfvbench.py +++ b/test/test_nfvbench.py @@ -623,10 +623,12 @@ def get_dummy_tg_config(chain_type, rate): 'no_arp': False, 'duration_sec': 1, 'interval_sec': 1, + 'pause_sec': 1, 'rate': rate, 'check_traffic_time_sec': 200, 'generic_poll_sec': 2, 'measurement': {'NDR': 0.001, 'PDR': 0.1, 'load_epsilon': 0.1}, + 'l2_loopback': False }) def get_traffic_client(): @@ -40,6 +40,6 @@ show-source = True #H404: multi line docstring should start without a leading new line #H405: multi line docstring summary not separated with an empty line #H904: Wrap long lines in parentheses instead of a backslash -ignore = E123,E125,H803,E302,E303,H104,H233,H236,H302,H404,H405,H904 +ignore = E123,E125,H803,E302,E303,H104,H233,H236,H302,H404,H405,H904,D102,D100,D107 builtins = _ exclude=venv,.git,.tox,dist,doc,*openstack/common*,*lib/python*,*egg,build,dib-venv |